We are building a fully-hosted platform for configuring, optimizing, and analyzing Calibrated Models. We believe that anyone from Data Scientists to Product Managers should be able to use Machine Learning systems that leverage their domain expertise while maintaining control and understanding over the predictions that their models make.

SOTAI

sotai

What if your avatar in the Genies Ecosystem encoded not only your visual representation but also who you are – your personality, quirks, and all? The Traits Framework at Genies is our approach to bringing multi-layered depth and dynamism to your digital identity: depicting not just what you look like, but also describing what your personality is like. Why do we want to do this? Because we think your online persona should grow and change just like you do in real life. It's all about making the digital world more personal and responsive to who you are. There’s nowhere else where you can be part of an ecosystem that not only comes with amazing high-fidelity avatars but also a high-fidelity description of the inner self and personality of someone in the digital world. We describe a little bit of this in our previous , but we’d like to dive deeper into our methodology so you can understand how we’ve designed our Traits Framework and how it describes your digital avatar’s inner self. It's like giving your digital self its own unique flavor. blog post The Origins of Our Traits Framework We had two initial requirements for our Traits Framework that our product team set out as goals It must be comprehensive enough to describe a person relatively accurately but also uniquely. It must be simple & straightforward enough for developers to build innovative custom experiences based on your unique Trait profile. Based on these two requirements, we knew we had to bring in experts to do proper discovery work in building out the foundation of Traits. Therefore, we brought in a team of psychologists to work directly with our product team in building the foundation of our innovative framework. The psychologists helped build a clinical and scientific foundation for our framework and helped audit our creative process to ensure rigor in design. The Game Designer helped provide a developer perspective in architecting this framework and is now developing flagship examples that will make it straightforward for a new genre of personality-based Experiences to be built by anyone building within the Genies Ecosystem. We first worked with the psychologists to pull inspiration and research from other personality systems (Jungian Archetypes, the 7 Chakra System, 5 elements in Chinese medicine, and the 9 Enneagram Personality Types), but in parallel, we also looked to innovate in creating our own structure that would work specifically for our ecosystem where we’re able to have a bit of creative leeway to make personality profiles much more fun and engaging especially as we need to pair them with Experiences. Eventually, our focus narrowed to combining the 7 Chakras System with a gaming approach of multi-layered Traits (i.e., FIFA’s DEF or “Defensive” stat is actually a composition of multiple sub-traits each with their own values), forming the bedrock for the current Traits Framework. Clinical judgment from psychologists and qualitative research methods helped us to distill each of the 7 Chakras into 7 categories, each with three descriptive Traits that uniquely describe a different facet of that respective category. We then worked with the psychologists to generate a list of 21 Traits in total (3 Traits per category, 7 categories) making sure that we don’t generate so many Traits that it’s impossible for developers to build against, but making sure we have enough variety in Traits such that every individual can eventually build up a unique Trait DNA profile. As seen in the graphic above, these 21 Traits can also be beautifully grouped by “Mind”, “Gut”, and “Heart” type Traits where Mind are the more intellectual Traits, Gut Traits are the instinctive Traits, and Heart Traits are the emotional Traits. We’ve also color-coded each category based on the chakra system to give each personality type more expression and something for both users and developers to hold on to. As we started to leverage this framework to create custom personality-tailored experiences, our Game Designer gave us valuable feedback to simplify the system as much as possible to make the development experience as easy for devs as possible in the near term. Thus, we then worked to focus on 7 Traits to start off with - one from each category that the psychologists felt epitomized each category the best and was most accessible to users. We are not removing the remaining 14 Traits; eventually, we will want to layer in the two additional Traits within each category as we expand our Framework to create better fidelity. These are the final starting 7 Traits that we ended up with along with their illustrative Trait stones that can be gained by users within the Genies Party ecosystem. This personality Trait dimension presents a unique opportunity for users to learn about themselves from spiritual, psychological, somatic, and material perspectives. It’s an invitation to reflect on who you are, who you want to be, and to see aspects of yourself reflected in the Genies ecosystem. This is not the end though! As we build out the Genies Party ecosystem, there will not only be these Foundational 21 personality Traits but also many other Traits added that will gradually create more fidelity as more complex Traits grow from these Foundational Traits. Some of these you can start to experience in our ! personality quiz Additionally, we are looking to expand and test additional Traits like Physical Traits, which are useful for games within the Genies Party ecosystem as developers may find these Traits easier to understand. The first physical Traits that you will be able to gain for your avatar in Genies Party are “Strength”, “Stamina”, and “Speed.” How Will a Genies Party User See These Traits? As a user builds out their Trait profile in Genies Party, they will be able to collect and build on the Trait stones that correspond to their personality, with their most prominent Traits shown closer to their Avatar. There’s no “best” Trait value or combination for users, you are your own unique person and every profile will be a relatively unique combination of scores for each Trait. It’s like a Trait DNA that’s created for each user and constantly evolving based on your actions within Genies Party. Your Trait profile will be developed through what you say and do in Genies Party. Every conversation and every experience could shape your profile based on your behavior. It’s also fun seeing how your Trait profile could evolve and change over time, maybe even to characteristics you wouldn’t have expected! Don’t worry, if you don’t like your Trait profile, you can always reset your Traits and start over just as you would be able to completely change up your avatar appearance. Because this Trait profile is a result of your actions taken in the digital world, it’ll truly become a horoscope of your personality. Through Genies Party, we’d also capture insights about your inner self that you may not have been able to realize through any other platform, allowing you to compare your true self to friends and Creators. “Am I more similar to my friend John or Claire?” “Would I be compatible with Timothee Chalamet?” Finally, the personalized experiences brought to life by your Traits in Genies Party are set to redefine avatar-based gaming. They unlock a new realm of customized games that not only enhance how you can socialize your identity but also add a whole new tailored-to-you dimension of multiplayer interactions. We’ll touch on this in the next section along with some challenges of balance that we’ve thought of between Traits. It’s important to note that although the creation of this Framework is rooted in empirical and scientific work, our aim is to still make it fun with somewhat mystical/metaphysical leanings, rather than a more standard personality profile measure like the Myers-Briggs or the Big Five personality Traits. How Will Custom Experiences Be Built Against These Traits? Devs should generally not be using Traits to modify the difficulty of a game. Our recommendation for types of ways 3rd party devs can develop against Traits are as such: - Through cosmetic changes in avatar, environment, and VFX, devs can provide unique Trait-based experiences that don’t directly impact game mechanics. This rewards players with a sense of achievement without disrupting the game flow. Cosmetic - Through narrative changes in game flow, story, or progression, a user’s Traits can change the way NPCs interact or how the game progresses in story. Narrative - Through specific Trait values, a user may be given a shortcut through a game experience. Shortcuts/Paths If devs have created multiple archetypes of characters in their game, Traits can be used to assign an archetype to that player. Archetypes - - Things are the smart interoperable 3D objects of the Genies Ecosystem. Traits can be attached to Things developers create, adding another depth and dynamism to not just avatars, but also the objects that you carry around. Things For example, if users have a high Calm Trait in a Tamagotchi-type game, they may be more likely to find calmer pets to raise. Similarly, if a user picks a calmer pet to raise within that same game, their Calm Trait value may increase. Another example is if in a racing game, a user comes in with a specific Trait, they may get 3 race bonuses to pick from based on their specific Trait, whereas someone coming in with a different Trait may be selecting from 3 different race bonuses. Cosmetically, you may notice that Caring Trait users may have a pink heart-particle aura and Ambitious Trait users may have glowing red exhausts. We do not want users to use Traits for gating or power level purposes within their games to avoid balance issues. It’s also important not to only have direct player competition within games based on Traits to not have to contend with “which Trait is better” type of questions for each game. If you have concerns about being assigned a Trait profile that doesn't resonate with you, we have a thoughtful solution: Things x Traits. Envision this: each item your avatar equips carries specific Traits. This allows you the flexibility to customize your Trait profile to suit various experiences, ensuring that each interaction can be shaped by your choices and the items you choose to equip, giving you the power to own your gaming experience. We’ll touch on our innovative Things Framework and the intricacies of that system another time. How do Developers Access Traits in Our Developer Kit? We've integrated an API into our devkit scripting, allowing developers to access Trait values ranging from -100 to 100. While the exact number of a Trait isn't directly shown, this range simplifies the application of Traits in-game logic. For instance, a developer can use a formula like `if(max(calm, 0) * RNG > 50) success! Reading Traits: Else failure to apply the positive aspect of the calm Trait in a random number generator (RNG) check. This approach simplifies complex calculations, making the development process more efficient. Our API allows developers to adjust Traits in terms of Large, Medium, and Small increments, either positively or negatively. For example, a period of low actions per minute (APM) with successful outcomes might prompt a Positive Medium adjustment to the Calm Trait. Writing Traits: Conversely, a high APM with few successes could lead to a Negative Medium adjustment. We're also exploring a 'batching' API for more granular games, allowing developers to accumulate Trait changes and send them in batched API calls. This ensures smoother and more efficient Trait management in our devkit. To further simplify Trait adjustments, we're considering offering developers various controllers that they can easily integrate and configure. These controllers would respond to different gameplay aspects, like APM and success rates, to automatically adjust Traits. By simplifying Trait management through intuitive APIs and developer-friendly tools, our goal is to empower developers to create a completely new genre of more engaging and dynamic games that are effectively customized to each user's Trait profile. Stay tuned for more updates on these exciting developments in our developer toolkit, and look forward to our first flagship Trait-based games built internally by Genies and by 3rd party devs in our Developer Incubator. If you’re interested in building on our innovative Traits Framework and being an early creator within the Genies Ecosystem, apply for our today. You can also learn about what we’re doing with AI Avatars on our . Developer Incubator website . Developer Incubator todayDeveloper Incubator todayDeveloper Incubator today

A Deep Dive on Digital Identity: Why You Need to Check Out Genies Traits Framework

ডিজিটাল আইডেন্টিটি সম্পর্কে গভীর ডুব: কেন আপনাকে জেনেসিসের বৈশিষ্ট্য কাঠামো পরীক্ষা করতে হবে

Ein tiefer Einblick in die digitale Identität: Warum Sie sich das Traits Framework von Genesis ansehen sollten

Una inmersión profunda en la identidad digital: por qué es necesario consultar el marco de rasgos de Génesis

Une analyse approfondie de l'identité numérique : pourquoi vous devez consulter le cadre de traits de Genesis

डिजिटल पहचान पर एक गहन जानकारी: आपको जेनेसिस के लक्षण ढांचे की जांच करने की आवश्यकता क्यों है

デジタル アイデンティティの詳細: Genesis の特性フレームワークをチェックする必要がある理由

디지털 신원에 대한 심층 분석: Genesis의 특성 프레임워크를 확인해야 하는 이유

Um mergulho profundo na identidade digital: por que você precisa verificar a estrutura de características do Genesis

Глубокое погружение в цифровую идентичность: почему вам нужно ознакомиться с системой черт Genesis

Dijital Kimliğe Derin Bir Bakış: Genesis'in Özellikler Çerçevesine Neden Göz Atmanız Gerekiyor?

Tìm hiểu sâu về nhận dạng kỹ thuật số: Tại sao bạn cần kiểm tra Khung đặc điểm của Genesis

深入探讨数字身份：为什么您需要查看 Genesis 的特征框架

In today's edition of "Behind the Startup," we delve deep into the world of blockchain, touching upon everything from its inception to the impact of regulatory pressures on the crypto realm. We're privileged to have Louis Tsu, the CEO & CTO of the Venom Foundation blockchain and CEO of Venom Africa, join us. With a rich history in technology and a deep understanding of macroeconomics, Louis offers a unique lens through which we can understand the complex intricacies of the rapidly evolving digital currency landscape. A Sovereign-Level Scalability and the Game-Changing Regulation from Abu Dhabi Ishan Pandey: Hi Louis, welcome to our series “Behind the Startup.” Please tell us about yourself and the journey that led you to become the CEO & CTO of the Venom Foundation blockchain and CEO of Venom Africa? In 1979 I was fixing Black & White TV’s, and soldering circuit boards. From that point I’ve always had one foot in technology, ex-Apple, ex-TI silicon chip and founder of start-up’s. The other foot was in macro, economics and finance. So when I saw Bitcoin in 2017 and got wind of Venom in 2021, that was it. Louis Tsu: Using Venom blockchain to bring better finance and help people especially in developing markets was right up my street. My transition from CTO into CEO was very natural as I have all the tools in my pocket understanding deep technology and vast experience in visionary leadership. Ishan Pandey: Could you provide us some insights about the Venom Foundation blockchain? What makes it stand out compared to other blockchain platforms? Louis Tsu: Two fundamental aspects: The blockchain can scale at a sovereign level and maintain steady transaction speeds at low cost. It is regulated by the Abu Dhabi Global Markets. I could literally talk for hours on the tech. But at the highest level, this is my best summary. Ishan Pandey: The recent SEC crackdown on Binance, Coinbase, and other cryptos has been a major talking point. What is your opinion on this issue, particularly considering the SEC's belief that every crypto asset, except bitcoin, is a security? Being a security is not a crime, plus, it only matters if your in the USA and a) you are the originator of the crypto asset; b) the distribution reseller of the asset. All this means is more expensive paperwork. Louis Tsu: It is all about liquidity, crypto coins that are highly liquid have more robust price action. When the SEC delists a crypto asset, it becomes illiquid. Look at Bitcoin, it is ubiquitous and the most liquid crypto asset. Ishan Pandey: Do you see the recent SEC's regulatory actions as an attempt to establish control over a space that has been mostly unregulated or is this more of a 'ChokePoint 2.0' operation? Why? The SEC is very well-versed in crypto. I remember 3 or 4 years ago the SEC at the Senate were talking about how to manage atomic swaps. IMO the SEC is acting by design, in a coordinated manner to slow down the US crypto industry. Louis Tsu: It could be their opinion is such that well-established financial institutions could take over from the incumbents in crypto and do a better job, I do not know why the SEC is acting this way? Ishan Pandey: As a veteran entrepreneur with an extensive background in monetary policy and macroeconomics technology, how do you interpret the SEC's motivation behind these lawsuits and their potential implications for the US, and the world at large? The implications are good for the rest off the world and bad for the USA Louis Tsu: In some countries regulators are used as a hammer to smash an industry and in other countries the regulator will nurture that industry through collaboration. The PR the SEC get is overrated but very useful for blockchain infrastructure providers like ourselves. I get many questions from different sovereign nations. This spotlight shines on key topics such as the reliance on the US dollar and what role can blockchain play. In general, I see different countries looking to employ game theory to protect their national interests. Ishan Pandey: Given the legal pressures on cryptocurrency exchanges, how do you foresee the crypto industry responding to these recent events? If the USA SEC leadership or the government changes tactics and plays nice, the damage is already done. Web 3.0. and crypto firms are voting with their feet and it will be quite difficult to undo. Louis Tsu: More importantly, other countries are pressing ahead with sound regulation to attract this nascent industry into their jurisdiction. The actions of the SEC are like a beautiful sales brochure for all the other crypto-friendly regions across the world. You can't put that toothpaste back in the tube. Ishan Pandey: You've mentioned your belief that regulation is key for the crypto space, but not in the way it's currently being implemented. Could you elaborate on your vision for how regulation should be carried out? Indeed regulation will permit the crypto industry easy access to the mass population and hence global adoption. I have experienced sitting down with regulators and cooperating, exchanging ideas, learning from one another to make something better for everyone. Louis Tsu: Innovation will always move much faster than regulation, only through good communication channels and top down support for that industry can private enterprise and government bodies swiftly move in harmony. This is not new, I remember back in the day in London during the 1990’s I founded a company manufacturing digital TV products and within 18 months a big country in ASIA in fierce competition with Europe changed all their regulations on silicon chips in order to out compete the EU market for digital TV. Ishan Pandey: You noted the influx of massive institutions like Blackrock applying for digital asset licenses, and the apparent preparation of the SEC to regulate the US market. In this context, do you think traditional financial institutions are set to overtake original crypto shops that have been trying to play by the book? Traditional finance will not overtake or shall we say cannibalize crypto shops. The market is still in the changing room, we have not even gone on the track and stepped on the starting blocks yet. Louis Tsu: I’ve seen this playbook in the past. The entrance of the institutions will dramatically expand the overall market for everyone. Ishan Pandey: What do you think will be the implications for smaller, traditional crypto businesses if large financial institutions take a leading role in the crypto space? We are in a new dimension of software development, open-source and composability will move this market faster than Metcalf’s law. Smaller shops will build and they are in the best position to build on open-source. Institutions will have to adapt, I have no doubt they will, but they will not crush small shops. Louis Tsu: This is not going to be like when giant supermarkets killed small grocery stores, sure some will die of natural causes, but the shops that provide great value can easily shine in today’s modern world. It’s kind of like music, you think ‘that’s it’ no more new music riff’s or tunes can be created, we have all the best music in the world. Then suddenly a kid with a guitar writes a song that scythes through the charts. Ishan Pandey: Lastly, could you provide us with your outlook on the future of cryptocurrencies and blockchain technology, especially in light of the ongoing regulatory developments? The next bull run will not be the same as the last one. Many coins with no utility will dissipate. Capital will flow to value, and this will take a couple of years to get realized, in the short term we will see FOMO moving out the risk curve and irrational bets getting placed. But in the longer run, money always goes to the hardest asset and they will be the coins with intrinsic value. Louis Tsu: Blockchain will be a big winner because adoption is already underway and everyone needs the plumbing and wiring to build on. I see the regulation as putting a legitimate stamp on blockchain, such as Venom! : This author is an independent contributor publishing via our . Be it through direct compensation, media partnerships, or networking, the author has a vested interest in the company/ies mentioned in this story. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR Vested Interest Disclosure brand-as-author program

Louis Tsu Weighs In on the Future of US-Based Crypto Firms

Louis Tsu opina sobre el futuro de las criptoempresas con sede en EE. UU.

Louis Tsu se penche sur l'avenir des entreprises de cryptographie basées aux États-Unis

लुई त्सू यूएस-आधारित क्रिप्टो फर्मों के भविष्य पर विचार कर रहे हैं

ルイス・ツー氏、米国に本拠を置く仮想通貨企業の将来について考察

Louis Tsu avalia o futuro das empresas criptográficas sediadas nos EUA

Louis Tsu cân nhắc về tương lai của các công ty tiền điện tử có trụ sở tại Hoa Kỳ

Louis Tsu 评价美国加密货币公司的未来

FTX + SBF Chapter 11 (Bankruptcy) Court Filing by John J. Ray III, Nov 17, 2022 is part of . You can jump to any chapter in this filing . HackerNoon’s Legal PDF Series here Feature Image: Stable Diffusion AI, Prompt “Group email as private keys” III. ACTION TAKEN SINCE MR. BANKMAN-FRIED’S DEPARTURE G - Digital Asset Custody 65. The FTX Group did not keep appropriate books and records, or security controls, with respect to its digital assets. Mr. Bankman-Fried and Mr. Wang controlled access to digital assets of the main businesses in the FTX Group (with the exception of LedgerX, regulated by the CFTC, and certain other regulated and/or licensed subsidiaries). Unacceptable management practices included the use of an unsecured group email account as the root user to access confidential private keys and critically sensitive data for the FTX Group companies around the world, the absence of daily reconciliation of positions on the blockchain, the use of software to conceal the misuse of customer funds, the secret exemption of Alameda from certain aspects of FTX.com’s auto-liquidation protocol, and the absence of independent governance as between Alameda (owned 90% by Mr. Bankman-Fried and 10% by Mr. Wang) and the Dotcom Silo (in which third parties had invested). 66. The Debtors have located and secured only a fraction of the digital assets of the FTX Group that they hope to recover in these Chapter 11 Cases. The Debtors have secured in new cold wallets approximately $740 million of cryptocurrency that the Debtors believe is attributable to either the WRS, Alameda and/or Dotcom Silos. The Debtors have not yet been able to determine how much of this cryptocurrency is allocable to each Silo, or even if such an allocation can be determined. These balances exclude cryptocurrency not currently under the Debtors’ control as a result of (a) at least $372 million of unauthorized transfers initiated on the Petition Date, during which time the Debtors immediately began moving cryptocurrency into cold storage to mitigate the risk to the remaining cryptocurrency that was accessible at the time, (b) the dilutive ‘minting’ of approximately $300 million in FTT tokens by an unauthorized source after the Petition Date and (c) the failure of the co-founders and potentially others to identify additional wallets believed to contain Debtor assets. 67. In response, the Debtors have engaged forensic analysts to identify potential Debtor assets on the blockchain, cybersecurity professionals to identify the parties responsible for the unauthorized transactions on and after the Petition Date and investigators to begin the process of identifying what may be very substantial transfers of Debtor property in the days, weeks and months prior to the Petition Date. The Debtors’ team includes business, accounting, forensic, technical and legal resources that I believe are among the best in the world at these activities. It is my expectation that the Debtors will require assistance from the Court with respect to these matters as the investigation and these Chapter 11 Cases continue. 68. Although the investigation has only begun and must run its course, it is my view based on the information obtained to date, that many of the employees of the FTX Group, including some of its senior executives, were not aware of the shortfalls or potential commingling of digital assets. Indeed, I believe some of the people most hurt by these events are current and former employees and executives, whose personal investments and reputations have suffered. These are many of the same people whose work will be necessary to ensure the maximization of value for all stakeholders going forward. Continue Reading Here

There was no digital security at FTX; a group email was used as key to all assets

No había seguridad digital en FTX; se utilizó un correo electrónico de grupo como clave para todos los activos

Il n'y avait pas de sécurité numérique chez FTX ; un e-mail de groupe a été utilisé comme clé pour tous les actifs

एफटीएक्स पर कोई डिजिटल सुरक्षा नहीं थी; एक समूह ईमेल का उपयोग सभी संपत्तियों की कुंजी के रूप में किया गया था

FTX にはデジタル セキュリティはありませんでした。グループの電子メールがすべてのアセットのキーとして使用されました

Não havia segurança digital na FTX; um e-mail de grupo foi usado como chave para todos os ativos

Không có bảo mật kỹ thuật số tại FTX; một email nhóm đã được sử dụng làm khóa cho tất cả nội dung

FTX 没有数字安全；群组电子邮件被用作所有资产的关键

The Romance of Modern Mechanism by Archibald Williams is part of the HackerNoon Books Series. You can jump to any chapter in this book . CHAPTER XXI here A SELF-MOVING STAIRCASE At the American Exhibition, held in the Crystal Palace in 1902, there was shown a staircase which, on payment of a penny, transported any sufficiently daring person from the ground-floor to the gallery above. All that the experimenters had to do was to step boldly on, take hold of the balustrade, which moved at an equal pace with the stairs, and step off when the upper level was reached. The "escalator" (Latin = flight of stairs) hails from the United States, where it is proving a serious rival to the elevator. In principle, it is a continuously working lift, the slow travel of which is more than compensated by the fact that it is always available. The ordinary elevator is very useful in a large business or commercial house, where it saves the legs of people who, if they had to tramp up flight after flight of stairs, would probably not spend so much money as they would be ready to part with if their vertical travel from one floor to another was entirely free of effort. But the ordinary lift is, like a railway, intermittent. We all know what it means to stand at the grille and watch the cage slide downwards on its journey of, perhaps, four floors, when we want to go to a floor higher up. Rather than face the delay we use our legs. scalae Theoretically, therefore, a large emporium should contain at least two lifts. If the number be further increased, the would-be passenger will have a still better chance of getting off at once. Thus at the station of the Central London Railway we have to wait but a very few seconds before a grille is thrown back and an attendant invites us to "Hurry up there, please!" Yet there is delay while the cage is being filled. The actual journey occupies but a small fraction of the time which elapses between the moment when the first passenger enters the lift at the one end of the trip and the moment when the last person leaves it at the other end. In a building where the lift stops every fifteen feet or so to take people on or put them off, the waste of time is still more accentuated. The escalator is always ready. You step on and are transported one stage. A second staircase takes you on at once if you desire it. There is no delay. Furthermore, the room occupied by a single escalator is much less than that occupied by the number of lifts required to give anything like an equally efficient service. In large American stores, then, it is coming into favour, and also on the Manhattan Elevated Railway of New York. When once the little nervousness accompanying the first use has worn off, it eclipses the lift. A writer in Cassier's Magazine says: "In one large retail store during the holiday season more than 6,000 persons per hour have been carried upon the escalator for five hours of the day, and the aggregate for an entire day is believed to be 50,000. In the same store on an ordinary day the passengers alighting at the second floor from the eight large lifts, which run from the basement to the fifth floor, were counted, likewise the number at the escalator. This latter was found to be 859 per cent. of the number delivered by the eight lifts. In another establishment, in a very busy hour, the number taken from the first floor by the escalator was four times the number taken from the first floor by the fourteen lifts, which were running at their maximum capacity. To the merchant this spells opportunity for business. "The experience at the Twenty-third Street and Sixth Avenue station of the Manhattan Elevated Railway in New York, during a recent shut-down of the escalator, which has been in service for some time, is interesting as showing the attitude of the public, of which many millions have been carried by the installation during the several years of its operation. The daily traffic receipts of this station for a period beginning several weeks before the shut-down and extending as many after, for the years 1903 and 1902, and receipts of the adjacent stations for the same period were carefully plotted ... and the loss area during the period of shut-down was determined. The loss area was found to embrace 64,645 fares. It was, furthermore, daily a matter of observation that numbers of people, finding that the escalator was not running, refused to climb the stairs, and turned away from the station. "In the case of a great store, the escalator may be constructed as one continuous machine, with landings at each floor, and so arranged that steps which carry passengers up may perform a like service in carrying others down; or separate machines may be installed in various locations affording the best opportunity for displaying merchandise to the customer who may be proceeding from the lower to the upper floor. In the case of a six-storey building so equipped with escalator service in both directions, or in all ten escalator flights, it is obvious that the facilities are equal to an impossible number of elevators; and as facility of access has a direct bearing upon opportunities for business, it may well be argued that the relative value, measured by rent, of the main and upper floors is greatly changed." Each step in a staircase has two parts—the "tread" or horizontal board on which the foot is placed, and the vertical "riser" which acts both as a support to the tread above and also prevents the foot from slipping under the tread. In the escalator each tread is attached rigidly to its riser, and the two together form an independent unit. For the convenience of passengers in stepping on or off at the upper and lower landings, the treads in these places are all in the same horizontal plane. As they approach the incline the risers gradually appear, and the treads separate vertically. At the top of the incline the process is gradually reversed, the risers disappearing until the treads once more form a horizontal belt. The means of effecting this change is most ingenious. Each tread and its riser is carried on a couple of vertical triangular brackets, one at each side of the staircase. The base of the bracket is uppermost, to engage with the tread, and its apex has a hole through which passes a transverse bar, which in its central part forms a pin in the link-chain by which power is transmitted to the escalator. Naturally, the step would tip over. This is prevented by a yoke attached to each end of the bar, at right angles to it and parallel to the tread. The yoke has at each extremity a small wheel running on its own rail—there being two rails for each side of the staircase. Since step, brackets, bar, and yoke are all rigidly joined together, the step is unable to leave the horizontal, but its relation to the steps above and below is determined by the arrangement of the rails on which the yoke wheels run. When these are in the same plane, all the yokes, and consequently the treads, will also be in the same plane. But at the incline, where the inner rail gradually sinks lower than its fellow, the front wheel of one tread is lower than the front wheel of the next, and the risers appear. It may be added that, owing to the double track at each side of the staircase, the back wheel of one tread does not interfere with the front wheel of that below; and that on the level they come abreast without jostling, as the yoke is bent. The chain, of which the step-bars form pins, travels under the centre of the staircase. It is made up of links eighteen inches long, having, in addition to the bars, a number of steel cross-pins 112 inches in diameter, their axes three inches apart, so that the chain as a whole has a three-inch "pitch." The hubs of the links are bushed with bronze, and have a graphite "inlay," which makes them self-lubricating. Every joint is turned to within 11,000 inch of absolute accuracy. The tracks are of steel and hardwood, insulated from the ironwork which supports them by sheets of rubber. The wheels are so constructed as to be practically noiseless, so that as a whole the escalator works very quietly. "It has been observed," says the authority already quoted, "that beginners take pains to step upon a single tread, and that after a little experience no attention whatever is given to the footing, owing to the facility of adapting oneself to the situation. The upper landing is somewhat longer, thereby affording an interval for stepping off at either side of sufficient duration to meet the requirements of the aged and infirm. The sole function of the travelling landing is to provide a time interval to meet the requirements of the slowest-acting passenger, and not of the alert. The terminal of the exit landing, be it top or bottom (for the escalator operates equally well for either ascent or descent), is a barrier, called the shunt, of which the lower member travels horizontally in a plane oblique to the direction of movement of the steps, and at a speed proportionately greater, thereby imparting a right-angle resultant to the person or obstacle on the step which may come in contact with the shunt. By reason of this resultant motion, the person or obstacle is gently pushed off the end of the step upon the floor, without shock or injury in the slightest degree. The motion of the escalator is so smooth and constant that it does not interpose the least obstacle to the free movement of the passenger, who may walk in either direction or assume any attitude to the same degree as upon a stationary staircase." At Cleveland, U.S.A., there has been erected a rolling roadway, consisting of an inclined endless belt and platform made of planks eight feet long, placed transversely across the roadway. The timbers are fastened together in trucks of two planks each, adjoining trucks being joined by heavy links to form a moving roadway, which runs on 4,000 small wheels. At each end the roadway, which is continuous, passes round enormous rollers. Its total length is 420 feet, and the rise 65 feet. Four electric motors placed at regular intervals along its length, and all controlled by one man at the head of the incline, drive it at three miles an hour. It can accommodate six wagons at a time. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Archibald Williams (2014). The Romance of Modern Mechanism. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/46094/pg46094-images.html. This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

A SELF-MOVING STAIRCASE

The Big Four by Agatha Christie, is part of the HackerNoon Books Series. You can jump to any chapter in this book here . THE WOMAN ON THE STAIRS VI. THE WOMAN ON THE STAIRS That was all that could be elicited from Mrs. Halliday. We hurried back to London, and the following day saw us en route for the Continent. With rather a rueful smile, Poirot observed:—- "This Big Four, they make me to bestir myself, mon ami. I run up and down, all over the ground, like our old friend 'the human foxhound.'" "Perhaps you'll meet him in Paris," I said, knowing that he referred to a certain Giraud, one of the most trusted detectives of the Sureté, whom he had met on a previous occasion. Poirot made a grimace. "I devoutly hope not. He loved me not, that one." "Won't it be a very difficult task?" I asked. "To find out what an unknown Englishman did on an evening two months ago?" "Very difficult, mon ami. But, as you know well, difficulties rejoice the heart of Hercule Poirot." "You think the Big Four kidnapped him?" Poirot nodded. Our inquiries necessarily went over old ground, and we learnt little to add to what Mrs. Halliday had already told us. Poirot had a lengthy interview with Professor Bourgoneau, during which he sought to elicit whether Halliday had mentioned any plan of his own for the evening, but we drew a complete blank. Our next source of information was the famous Madame Olivier. I was quite excited as we mounted the steps of her villa at Passy. It has always seemed to me extraordinary that a woman should go so far in the scientific world. I should have thought a purely masculine brain was needed for such work. The door was opened by a young lad of seventeen or thereabouts, who reminded me vaguely of an acolyte, so ritualistic was his manner. Poirot had taken the trouble to arrange our interview beforehand, as he knew Madame Olivier never received any one without an appointment, being immersed in research work most of the day. We were shown into a small salon, and presently the mistress of the house came to us there. Madame Olivier was a very tall woman, her tallness accentuated by the long white overall she wore, and a coif like a nun's that shrouded her head. She had a long pale face, and wonderful dark eyes that burnt with a light almost fanatical. She looked more like a priestess of old than a modern Frenchwoman. One cheek was disfigured by a scar, and I remembered that her husband and co-worker had been killed in an explosion in the laboratory three years before, and that she herself had been terribly burned. Ever since then she had shut herself away from the world, and plunged with fiery energy into the work of scientific research. She received us with cold politeness. "I have been interviewed by the police many times, messieurs. I think it hardly likely that I can help you, since I have not been able to help them." "Madame, it is possible that I shall not ask you quite the same questions. To begin with, of what did you talk together, you and M. Halliday?" She looked a trifle surprised. "But of his work! His work—and also mine." "Did he mention to you the theories he had embodied recently in his paper read before the British Association?" "Certainly he did. It was chiefly of those we spoke." "His ideas were somewhat fantastic, were they not?" asked Poirot carelessly. "Some people have thought so. I do not agree." "You considered them practicable?" "Perfectly practicable. My own line of research has been somewhat similar, though not undertaken with the same end in view. I have been investigating the gamma rays emitted by the substance usually known as Radium C., a product of Radium emanation, and in doing so I have come across some very interesting magnetical phenomena. Indeed, I have a theory as to the actual nature of the force we call magnetism, but it is not yet time for my discoveries to be given to the world. Mr. Halliday's experiments and views were exceedingly interesting to me." Poirot nodded. Then he asked a question which surprised me. "Madame, where did you converse on these topics. In here?" "No, monsieur. In the laboratory." "May I see it?" "Certainly." She led the way to the door from which she had entered. It opened on a small passage. We passed through two doors and found ourselves in the big laboratory, with its array of beakers and crucibles and a hundred appliances of which I did not even know the names. There were two occupants, both busy with some experiment. Madame Olivier introduced them. "Mademoiselle Claude, one of my assistants." A tall, serious-faced young girl bowed to us. "Monsieur Henri, an old and trusted friend." The young man, short and dark, bowed jerkily. Poirot looked round him. There were two other doors besides the one by which we had entered. One, madame explained, led into the garden, the other into a smaller chamber also devoted to research. Poirot took all this in, then declared himself ready to return to the salon. "Madame, were you alone with M. Halliday during your interview?" "Yes, monsieur. My two assistants were in the smaller room next door." "Could your conversation be overheard—by them or any one else?" Madame reflected, then shook her head. "I do not think so. I am almost sure it could not. The doors were all shut." "Could any one have been concealed in the room?" "There is the big cupboard in the corner—but the idea is absurd." "Pas tout à fait, madame. One thing more: did M. Halliday make any mention of his plans for the evening?" "He said nothing whatever, monsieur." "I thank you, madame, and I apologise for disturbing you. Pray do not trouble—we can find our way out." We stepped out into the hall. A lady was just entering the front door as we did so. She ran quickly up the stairs, and I was left with an impression of the heavy mourning that denotes a French widow. "A most unusual type of woman, that," remarked Poirot, as we walked away. "Madame Olivier? Yes, she—" "Mais non, not Madame Olivier. Cela va sans dire! There are not many geniuses of her stamp in the world. No, I referred to the other lady—the lady on the stairs." "I didn't see her face," I said, staring. "And I hardly see how you could have done. She never looked at us." "That is why I said she was an unusual type," said Poirot placidly. "A woman who enters her home—for I presume that it is her home since she enters with a key—and runs straight upstairs without even looking at two strange visitors in the hall to see who they are, is a very unusual type of woman—quite unnatural, in fact. Mille tonnerres! what is that?" He dragged me back—just in time. A tree had crashed down on to the side walk, just missing us. Poirot stared at it, pale and upset. "It was a near thing that! But clumsy, all the same—for I had no suspicion—at least hardly any suspicion. Yes, but for my quick eyes, the eyes of a cat, Hercule Poirot might now be crushed out of existence—a terrible calamity for the world. And you, too, mon ami—though that would not be such a national catastrophe." "Thank you," I said coldly. "And what are we going to do now?" "Do?" cried Poirot. "We are going to think. Yes, here and now, we are going to exercise our little gray cells. This M. Halliday now, was he really in Paris? Yes, for Professor Bourgoneau, who knows him, saw and spoke to him." "What on earth are you driving at?" I cried. "That was Friday morning. He was last seen at eleven Friday night—but was he seen then?" "The porter—" "A night porter—who had not previously seen Halliday. A man comes in, sufficiently like Halliday—-we may trust Number Four for that—asks for letters, goes upstairs, packs a small suit-case, and slips out the next morning. Nobody saw Halliday all that evening—no, because he was already in the hands of his enemies. Was it Halliday whom Madame Olivier received? Yes, for though she did not know him by sight, an imposter could hardly deceive her on her own special subject. He came here, he had his interview, he left. What happened next?" Seizing me by the arm, Poirot was fairly dragging me back to the villa. "Now, mon ami, imagine that it is the day after the disappearance, and that we are tracking footprints. You love footprints, do you not? See—here they go, a man's, Mr. Halliday's.... He turns to the right as we did, he walks briskly—ah! other footsteps following behind—very quickly—small footsteps, a woman's. See, she catches him up—a slim young woman, in a widow's veil. 'Pardon, monsieur, Madame Olivier desires that I recall you.' He stops, he turns. Now where would the young woman take him? She does not wish to be seen walking with him. Is it coincidence that she catches up with him just where a narrow alleyway opens, dividing two gardens? She leads him down it. 'It is shorter this way, monsieur.' On the right is the garden of Madame Olivier's villa, on the left the garden of another villa—and from that garden, mark you, the tree fell—so nearly on us. Garden doors from both open on the alley. The ambush is there. Men pour out, overpower him, and carry him into the strange villa." "Good gracious, Poirot," I cried, "are you pretending to see all this?" "I see it with the eyes of the mind, mon ami. So, and only so, could it have happened. Come, let us go back to the house." "You want to see Madame Olivier again?" Poirot gave a curious smile. "No, Hastings, I want to see the face of the lady on the stairs." "Who do you think she is, a relation of Madame Olivier's?" "More probably a secretary—and a secretary engaged not very long ago." The same gentle acolyte opened the door to us. "Can you tell me," said Poirot, "the name of the lady, the widow lady, who came in just now?" "Madame Veroneau? Madame's secretary?" "That is the lady. Would you be so kind as to ask her to speak to us for a moment." The youth disappeared. He soon reappeared. "I am sorry. Madame Veroneau must have gone out again." "I think not," said Poirot quietly. "Will you give her my name, M. Hercule Poirot, and say that it is important I should see her at once, as I am just going to the Prefecture." Again our messenger departed. This time the lady descended. She walked into the salon. We followed her. She turned and raised her veil. To my astonishment I recognised our old antagonist, the Countess Rossakoff, a Russian countess, who had engineered a particularly smart jewel robbery in London. "As soon as I caught sight of you in the hall, I feared the worst," she observed plaintively. "My dear Countess Rossakoff—" She shook her head. "Inez Veroneau now," she murmured. "A Spaniard, married to a Frenchman. What do you want of me, M. Poirot? You are a terrible man. You hunted me from London. Now, I suppose, you will tell our wonderful Madame Olivier about me, and hunt me from Paris? We poor Russians, we must live, you know." "It is more serious than that, madame," said Poirot, watching her. "I propose to enter the villa next door, and release M. Halliday, if he is still alive. I know everything, you see." I saw her sudden pallor. She bit her lip. Then she spoke with her usual decision. "He is still alive—but he is not at the villa. Come, monsieur, I will make a bargain with you. Freedom for me—and M. Halliday, alive and well, for you." "I accept," said Poirot. "I was about to propose the same bargain myself. By the way, are the Big Four your employers, madame?" Again I saw that deathly pallor creep over her face, but she left his question unanswered. Instead, "You permit me to telephone?" she asked, and crossing to the instrument she rang up a number. "The number of the villa," she explained, "where our friend is now imprisoned. You may give it to the police—the nest will be empty when they arrive. Ah! I am through. Is that you, André? It is I, Inez. The little Belgian knows all. Send Halliday to the hotel, and clear out." She replaced the receiver, and came towards us, smiling. "You will accompany us to the hotel, madame." "Naturally. I expected that." I got a taxi, and we drove off together. I could see by Poirot's face that he was perplexed. The thing was almost too easy. We arrived at the hotel. The porter came up to us. "A gentleman has arrived. He is in your rooms. He seems very ill. A nurse came with him, but she has left." "That is all right," said Poirot, "he is a friend of mine." We went upstairs together. Sitting in a chair by the window was a haggard young fellow who looked in the last stages of exhaustion. Poirot went over to him. "Are you John Halliday?" The man nodded. "Show me your left arm. John Halliday has a mole just below the left elbow." The man stretched out his arm. The mole was there. Poirot bowed to the countess. She turned and left the room. A glass of brandy revived Halliday somewhat. "My God!" he muttered. "I have been through hell—hell.... Those fiends are devils incarnate. My wife, where is she? What does she think? They told me that she would believe—would believe—" "She does not," said Poirot firmly. "Her faith in you has never wavered. She is waiting for you—she and the child." "Thank God for that. I can hardly believe that I am free once more." "Now that you are a little recovered, monsieur, I should like to hear the whole story from the beginning." Halliday looked at him with an indescribable expression. "I remember—nothing," he said. "What?" "Have you ever heard of the Big Four?" "Something of them," said Poirot dryly. "You do not know what I know. They have unlimited power. If I remain silent, I shall be safe—if I say one word—not only I, but my nearest and dearest will suffer unspeakable things. It is no good arguing with me. I know.... I remember—nothing." And, getting up, he walked from the room. Poirot's face wore a baffled expression. "So it is like that, is it?" he muttered. "The Big Four win again. What is that you are holding in your hand, Hastings?" I handed it to him. "The countess scribbled it before she left," I explained. He read it. "Au revoir.—I.V." "Signed with her initials—I.V. Just a coincidence, perhaps, that they also stand for Four. I wonder, Hastings, I wonder." About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Agatha Christie (2023). The Big Four. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/70114/pg70114-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at www.gutenberg.org , located at https://www.gutenberg.org/policy/license.html .

THE WOMAN ON THE STAIRS

At the Earth's Core by Edgar Rice Burroughs, is part of the HackerNoon Books Series. You can jump to any chapter in this book here . SLAVES SLAVES AS WE DESCENDED THE BROAD STAIRCASE WHICH led to the main avenue of Phutra I caught my first sight of the dominant race of the inner world. Involuntarily I shrank back as one of the creatures approached to inspect us. A more hideous thing it would be impossible to imagine. The all-powerful Mahars of Pellucidar are great reptiles, some six or eight feet in length, with long narrow heads and great round eyes. Their beak-like mouths are lined with sharp, white fangs, and the backs of their huge, lizard bodies are serrated into bony ridges from their necks to the end of their long tails. Their feet are equipped with three webbed toes, while from the fore feet membranous wings, which are attached to their bodies just in front of the hind legs, protrude at an angle of 45 degrees toward the rear, ending in sharp points several feet above their bodies. I glanced at Perry as the thing passed me to inspect him. The old man was gazing at the horrid creature with wide astonished eyes. When it passed on, he turned to me. "A rhamphorhynchus of the Middle Olitic, David," he said, "but, gad, how enormous! The largest remains we ever have discovered have never indicated a size greater than that attained by an ordinary crow." As we continued on through the main avenue of Phutra we saw many thousand of the creatures coming and going upon their daily duties. They paid but little attention to us. Phutra is laid out underground with a regularity that indicates remarkable engineering skill. It is hewn from solid limestone strata. The streets are broad and of a uniform height of twenty feet. At intervals tubes pierce the roof of this underground city, and by means of lenses and reflectors transmit the sunlight, softened and diffused, to dispel what would otherwise be Cimmerian darkness. In like manner air is introduced. Perry and I were taken, with Ghak, to a large public building, where one of the Sagoths who had formed our guard explained to a Maharan official the circumstances surrounding our capture. The method of communication between these two was remarkable in that no spoken words were exchanged. They employed a species of sign language. As I was to learn later, the Mahars have no ears, not any spoken language. Among themselves they communicate by means of what Perry says must be a sixth sense which is cognizant of a fourth dimension. I never did quite grasp him, though he endeavored to explain it to me upon numerous occasions. I suggested telepathy, but he said no, that it was not telepathy since they could only communicate when in each others' presence, nor could they talk with the Sagoths or the other inhabitants of Pellucidar by the same method they used to converse with one another. "What they do," said Perry, "is to project their thoughts into the fourth dimension, when they become appreciable to the sixth sense of their listener. Do I make myself quite clear?" "You do not, Perry," I replied. He shook his head in despair, and returned to his work. They had set us to carrying a great accumulation of Maharan literature from one apartment to another, and there arranging it upon shelves. I suggested to Perry that we were in the public library of Phutra, but later, as he commenced to discover the key to their written language, he assured me that we were handling the ancient archives of the race. During this period my thoughts were continually upon Dian the Beautiful. I was, of course, glad that she had escaped the Mahars, and the fate that had been suggested by the Sagoth who had threatened to purchase her upon our arrival at Phutra. I often wondered if the little party of fugitives had been overtaken by the guards who had returned to search for them. Sometimes I was not so sure but that I should have been more contented to know that Dian was here in Phutra, than to think of her at the mercy of Hooja the Sly One. Ghak, Perry, and I often talked together of possible escape, but the Sarian was so steeped in his lifelong belief that no one could escape from the Mahars except by a miracle, that he was not much aid to us—his attitude was of one who waits for the miracle to come to him. At my suggestion Perry and I fashioned some swords of scraps of iron which we discovered among some rubbish in the cells where we slept, for we were permitted almost unrestrained freedom of action within the limits of the building to which we had been assigned. So great were the number of slaves who waited upon the inhabitants of Phutra that none of us was apt to be overburdened with work, nor were our masters unkind to us. We hid our new weapons beneath the skins which formed our beds, and then Perry conceived the idea of making bows and arrows—weapons apparently unknown within Pellucidar. Next came shields; but these I found it easier to steal from the walls of the outer guardroom of the building. We had completed these arrangements for our protection after leaving Phutra when the Sagoths who had been sent to recapture the escaped prisoners returned with four of them, of whom Hooja was one. Dian and two others had eluded them. It so happened that Hooja was confined in the same building with us. He told Ghak that he had not seen Dian or the others after releasing them within the dark grotto. What had become of them he had not the faintest conception—they might be wandering yet, lost within the labyrinthine tunnel, if not dead from starvation. I was now still further apprehensive as to the fate of Dian, and at this time, I imagine, came the first realization that my affection for the girl might be prompted by more than friendship. During my waking hours she was constantly the subject of my thoughts, and when I slept her dear face haunted my dreams. More than ever was I determined to escape the Mahars. "Perry," I confided to the old man, "if I have to search every inch of this diminutive world I am going to find Dian the Beautiful and right the wrong I unintentionally did her." That was the excuse I made for Perry's benefit. "Diminutive world!" he scoffed. "You don't know what you are talking about, my boy," and then he showed me a map of Pellucidar which he had recently discovered among the manuscript he was arranging. "Look," he cried, pointing to it, "this is evidently water, and all this land. Do you notice the general configuration of the two areas? Where the oceans are upon the outer crust, is land here. These relatively small areas of ocean follow the general lines of the continents of the outer world. "We know that the crust of the globe is 500 miles in thickness; then the inside diameter of Pellucidar must be 7,000 miles, and the superficial area 165,480,000 square miles. Three-fourths of this is land. Think of it! A land area of 124,110,000 square miles! Our own world contains but 53,000,000 square miles of land, the balance of its surface being covered by water. Just as we often compare nations by their relative land areas, so if we compare these two worlds in the same way we have the strange anomaly of a larger world within a smaller one! "Where within vast Pellucidar would you search for your Dian? Without stars, or moon, or changing sun how could you find her even though you knew where she might be found?" The proposition was a corker. It quite took my breath away; but I found that it left me all the more determined to attempt it. "If Ghak will accompany us we may be able to do it," I suggested. Perry and I sought him out and put the question straight to him. "Ghak," I said, "we are determined to escape from this bondage. Will you accompany us?" "They will set the thipdars upon us," he said, "and then we shall be killed; but—" he hesitated—"I would take the chance if I thought that I might possibly escape and return to my own people." "Could you find your way back to your own land?" asked Perry. "And could you aid David in his search for Dian?" "Yes." "But how," persisted Perry, "could you travel to strange country without heavenly bodies or a compass to guide you?" Ghak didn't know what Perry meant by heavenly bodies or a compass, but he assured us that you might blindfold any man of Pellucidar and carry him to the farthermost corner of the world, yet he would be able to come directly to his own home again by the shortest route. He seemed surprised to think that we found anything wonderful in it. Perry said it must be some sort of homing instinct such as is possessed by certain breeds of earthly pigeons. I didn't know, of course, but it gave me an idea. "Then Dian could have found her way directly to her own people?" I asked. "Surely," replied Ghak, "unless some mighty beast of prey killed her." I was for making the attempted escape at once, but both Perry and Ghak counseled waiting for some propitious accident which would insure us some small degree of success. I didn't see what accident could befall a whole community in a land of perpetual daylight where the inhabitants had no fixed habits of sleep. Why, I am sure that some of the Mahars never sleep, while others may, at long intervals, crawl into the dark recesses beneath their dwellings and curl up in protracted slumber. Perry says that if a Mahar stays awake for three years he will make up all his lost sleep in a long year's snooze. That may be all true, but I never saw but three of them asleep, and it was the sight of these three that gave me a suggestion for our means of escape. I had been searching about far below the levels that we slaves were supposed to frequent—possibly fifty feet beneath the main floor of the building—among a network of corridors and apartments, when I came suddenly upon three Mahars curled up upon a bed of skins. At first I thought they were dead, but later their regular breathing convinced me of my error. Like a flash the thought came to me of the marvelous opportunity these sleeping reptiles offered as a means of eluding the watchfulness of our captors and the Sagoth guards. Hastening back to Perry where he pored over a musty pile of, to me, meaningless hieroglyphics, I explained my plan to him. To my surprise he was horrified. "It would be murder, David," he cried. "Murder to kill a reptilian monster?" I asked in astonishment. "Here they are not monsters, David," he replied. "Here they are the dominant race—we are the 'monsters'—the lower orders. In Pellucidar evolution has progressed along different lines than upon the outer earth. These terrible convulsions of nature time and time again wiped out the existing species—but for this fact some monster of the Saurozoic epoch might rule today upon our own world. We see here what might well have occurred in our own history had conditions been what they have been here. "Life within Pellucidar is far younger than upon the outer crust. Here man has but reached a stage analogous to the Stone Age of our own world's history, but for countless millions of years these reptiles have been progressing. Possibly it is the sixth sense which I am sure they possess that has given them an advantage over the other and more frightfully armed of their fellows; but this we may never know. They look upon us as we look upon the beasts of our fields, and I learn from their written records that other races of Mahars feed upon men—they keep them in great droves, as we keep cattle. They breed them most carefully, and when they are quite fat, they kill and eat them." I shuddered. "What is there horrible about it, David?" the old man asked. "They understand us no better than we understand the lower animals of our own world. Why, I have come across here very learned discussions of the question as to whether gilaks, that is men, have any means of communication. One writer claims that we do not even reason—that our every act is mechanical, or instinctive. The dominant race of Pellucidar, David, have not yet learned that men converse among themselves, or reason. Because we do not converse as they do it is beyond them to imagine that we converse at all. It is thus that we reason in relation to the brutes of our own world. They know that the Sagoths have a spoken language, but they cannot comprehend it, or how it manifests itself, since they have no auditory apparatus. They believe that the motions of the lips alone convey the meaning. That the Sagoths can communicate with us is incomprehensible to them. "Yes, David," he concluded, "it would entail murder to carry out your plan." "Very well then, Perry." I replied. "I shall become a murderer." He got me to go over the plan again most carefully, and for some reason which was not at the time clear to me insisted upon a very careful description of the apartments and corridors I had just explored. "I wonder, David," he said at length, "as you are determined to carry out your wild scheme, if we could not accomplish something of very real and lasting benefit for the human race of Pellucidar at the same time. Listen, I have learned much of a most surprising nature from these archives of the Mahars. That you may appreciate my plan I shall briefly outline the history of the race. "Once the males were all-powerful, but ages ago the females, little by little, assumed the mastery. For other ages no noticeable change took place in the race of Mahars. It continued to progress under the intelligent and beneficent rule of the ladies. Science took vast strides. This was especially true of the sciences which we know as biology and eugenics. Finally a certain female scientist announced the fact that she had discovered a method whereby eggs might be fertilized by chemical means after they were laid—all true reptiles, you know, are hatched from eggs. "What happened? Immediately the necessity for males ceased to exist—the race was no longer dependent upon them. More ages elapsed until at the present time we find a race consisting exclusively of females. But here is the point. The secret of this chemical formula is kept by a single race of Mahars. It is in the city of Phutra, and unless I am greatly in error I judge from your description of the vaults through which you passed today that it lies hidden in the cellar of this building. "For two reasons they hide it away and guard it jealously. First, because upon it depends the very life of the race of Mahars, and second, owing to the fact that when it was public property as at first so many were experimenting with it that the danger of over-population became very grave. "David, if we can escape, and at the same time take with us this great secret what will we not have accomplished for the human race within Pellucidar!" The very thought of it fairly overpowered me. Why, we two would be the means of placing the men of the inner world in their rightful place among created things. Only the Sagoths would then stand between them and absolute supremacy, and I was not quite sure but that the Sagoths owed all their power to the greater intelligence of the Mahars—I could not believe that these gorilla-like beasts were the mental superiors of the human race of Pellucidar. "Why, Perry," I exclaimed, "you and I may reclaim a whole world! Together we can lead the races of men out of the darkness of ignorance into the light of advancement and civilization. At one step we may carry them from the Age of Stone to the twentieth century. It's marvelous—absolutely marvelous just to think about it." "David," said the old man, "I believe that God sent us here for just that purpose—it shall be my life work to teach them His word—to lead them into the light of His mercy while we are training their hearts and hands in the ways of culture and civilization." "You are right, Perry," I said, "and while you are teaching them to pray I'll be teaching them to fight, and between us we'll make a race of men that will be an honor to us both." Ghak had entered the apartment some time before we concluded our conversation, and now he wanted to know what we were so excited about. Perry thought we had best not tell him too much, and so I only explained that I had a plan for escape. When I had outlined it to him, he seemed about as horror-struck as Perry had been; but for a different reason. The Hairy One only considered the horrible fate that would be ours were we discovered; but at last I prevailed upon him to accept my plan as the only feasible one, and when I had assured him that I would take all the responsibility for it were we captured, he accorded a reluctant assent. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Edgar Rice Burroughs (1994). At the Earth's Core. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/123/pg123-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at www.gutenberg.org , located at https://www.gutenberg.org/policy/license.html.

WE DESCENDED THE BROAD STAIRCASE

Tarzan the Terrible by Edgar Rice Burroughs, is part of the HackerNoon Books Series. You can jump to any chapter in this book here . Blood-Stained Altars Blood-Stained Altars The entrance through which he caught his first glimpse of the interior was rather beautifully carved in geometric designs, and within the walls were similarly treated, though as he proceeded from one apartment to another he found also the figures of animals, birds, and men taking their places among the more formal figures of the mural decorator's art. Stone vessels were much in evidence as well as ornaments of gold and the skins of many animals, but nowhere did he see an indication of any woven fabric, indicating that in that respect at least the Ho-don were still low in the scale of evolution, and yet the proportions and symmetry of the corridors and apartments bespoke a degree of civilization. The way led through several apartments and long corridors, up at least three flights of stone stairs and finally out upon a ledge upon the western side of the building overlooking the blue lake. Along this ledge, or arcade, his guide led him for a hundred yards, to stop at last before a wide entrance-way leading into another apartment of the palace. Here Tarzan beheld a considerable concourse of warriors in an enormous apartment, the domed ceiling of which was fully fifty feet above the floor. Almost filling the chamber was a great pyramid ascending in broad steps well up under the dome in which were a number of round apertures which let in the light. The steps of the pyramid were occupied by warriors to the very pinnacle, upon which sat a large, imposing figure of a man whose golden trappings shone brightly in the light of the afternoon sun, a shaft of which poured through one of the tiny apertures of the dome. "Ko-tan!" cried Dak-lot, addressing the resplendent figure at the pinnacle of the pyramid. "Ko-tan and warriors of Pal-ul-don! Behold the honor that Jad-ben-Otho has done you in sending as his messenger his own son," and Dak-lot, stepping aside, indicated Tarzan with a dramatic sweep of his hand. Ko-tan rose to his feet and every warrior within sight craned his neck to have a better view of the newcomer. Those upon the opposite side of the pyramid crowded to the front as the words of the old warrior reached them. Skeptical were the expressions on most of the faces; but theirs was a skepticism marked with caution. No matter which way fortune jumped they wished to be upon the right side of the fence. For a moment all eyes were centered upon Tarzan and then gradually they drifted to Ko-tan, for from his attitude would they receive the cue that would determine theirs. But Ko-tan was evidently in the same quandary as they—the very attitude of his body indicated it—it was one of indecision and of doubt. The ape-man stood erect, his arms folded upon his broad breast, an expression of haughty disdain upon his handsome face; but to Dak-lot there seemed to be indications also of growing anger. The situation was becoming strained. Dak-lot fidgeted, casting apprehensive glances at Tarzan and appealing ones at Ko-tan. The silence of the tomb wrapped the great chamber of the throneroom of Pal-ul-don. At last Ko-tan spoke. "Who says that he is Dor-ul-Otho?" he asked, casting a terrible look at Dak-lot. "He does!" almost shouted that terrified noble. "And so it must be true?" queried Ko-tan. Could it be that there was a trace of irony in the chief's tone? Otho forbid! Dak-lot cast a side glance at Tarzan—a glance that he intended should carry the assurance of his own faith; but that succeeded only in impressing the ape-man with the other's pitiable terror. "O Ko-tan!" pleaded Dak-lot, "your own eyes must convince you that indeed he is the son of Otho. Behold his godlike figure, his hands, and his feet, that are not as ours, and that he is entirely tailless as is his mighty father." Ko-tan appeared to be perceiving these facts for the first time and there was an indication that his skepticism was faltering. At that moment a young warrior who had pushed his way forward from the opposite side of the pyramid to where he could obtain a good look at Tarzan raised his voice. "Ko-tan," he cried, "it must be even as Dak-lot says, for I am sure now that I have seen Dor-ul-Otho before. Yesterday as we were returning with the Kor-ul-lul prisoners we beheld him seated upon the back of a great GRYF. We hid in the woods before he came too near, but I saw enough to make sure that he who rode upon the great beast was none other than the messenger who stands here now." This evidence seemed to be quite enough to convince the majority of the warriors that they indeed stood in the presence of deity—their faces showed it only too plainly, and a sudden modesty that caused them to shrink behind their neighbors. As their neighbors were attempting to do the same thing, the result was a sudden melting away of those who stood nearest the ape-man, until the steps of the pyramid directly before him lay vacant to the very apex and to Ko-tan. The latter, possibly influenced as much by the fearful attitude of his followers as by the evidence adduced, now altered his tone and his manner in such a degree as might comport with the requirements if the stranger was indeed the Dor-ul-Otho while leaving his dignity a loophole of escape should it appear that he had entertained an impostor. "If indeed you are the Dor-ul-Otho," he said, addressing Tarzan, "you will know that our doubts were but natural since we have received no sign from Jad-ben-Otho that he intended honoring us so greatly, nor how could we know, even, that the Great God had a son? If you are he, all Pal-ul-don rejoices to honor you; if you are not he, swift and terrible shall be the punishment of your temerity. I, Ko-tan, King of Pal-ul-don, have spoken." "And spoken well, as a king should speak," said Tarzan, breaking his long silence, "who fears and honors the god of his people. It is well that you insist that I indeed be the Dor-ul-Otho before you accord me the homage that is my due. Jad-ben-Otho charged me specially to ascertain if you were fit to rule his people. My first experience of you indicates that Jad-ben-Otho chose well when he breathed the spirit of a king into the babe at your mother's breast." The effect of this statement, made so casually, was marked in the expressions and excited whispers of the now awe-struck assemblage. At last they knew how kings were made! It was decided by Jad-ben-Otho while the candidate was still a suckling babe! Wonderful! A miracle! and this divine creature in whose presence they stood knew all about it. Doubtless he even discussed such matters with their god daily. If there had been an atheist among them before, or an agnostic, there was none now, for had they not looked with their own eyes upon the son of god? "It is well then," continued the ape-man, "that you should assure yourself that I am no impostor. Come closer that you may see that I am not as are men. Furthermore it is not meet that you stand upon a higher level than the son of your god." There was a sudden scramble to reach the floor of the throne-room, nor was Ko-tan far behind his warriors, though he managed to maintain a certain majestic dignity as he descended the broad stairs that countless naked feet had polished to a gleaming smoothness through the ages. "And now," said Tarzan as the king stood before him, "you can have no doubt that I am not of the same race as you. Your priests have told you that Jad-ben-Otho is tailless. Tailless, therefore, must be the race of gods that spring from his loins. But enough of such proofs as these! You know the power of Jad-ben-Otho; how his lightnings gleaming out of the sky carry death as he wills it; how the rains come at his bidding, and the fruits and the berries and the grains, the grasses, the trees and the flowers spring to life at his divine direction; you have witnessed birth and death, and those who honor their god honor him because he controls these things. How would it fare then with an impostor who claimed to be the son of this all-powerful god? This then is all the proof that you require, for as he would strike you down should you deny me, so would he strike down one who wrongfully claimed kinship with him." This line of argument being unanswerable must needs be convincing. There could be no questioning of this creature's statements without the tacit admission of lack of faith in the omnipotence of Jad-ben-Otho. Ko-tan was satisfied that he was entertaining deity, but as to just what form his entertainment should take he was rather at a loss to know. His conception of god had been rather a vague and hazy affair, though in common with all primitive people his god was a personal one as were his devils and demons. The pleasures of Jad-ben-Otho he had assumed to be the excesses which he himself enjoyed, but devoid of any unpleasant reaction. It therefore occurred to him that the Dor-ul-Otho would be greatly entertained by eating—eating large quantities of everything that Ko-tan liked best and that he had found most injurious; and there was also a drink that the women of the Ho-don made by allowing corn to soak in the juices of succulent fruits, to which they had added certain other ingredients best known to themselves. Ko-tan knew by experience that a single draught of this potent liquor would bring happiness and surcease from worry, while several would cause even a king to do things and enjoy things that he would never even think of doing or enjoying while not under the magical influence of the potion, but unfortunately the next morning brought suffering in direct ratio to the joy of the preceding day. A god, Ko-tan reasoned, could experience all the pleasure without the headache, but for the immediate present he must think of the necessary dignities and honors to be accorded his immortal guest. No foot other than a king's had touched the surface of the apex of the pyramid in the throneroom at A-lur during all the forgotten ages through which the kings of Pal-ul-don had ruled from its high eminence. So what higher honor could Ko-tan offer than to give place beside him to the Dor-ul-Otho? And so he invited Tarzan to ascend the pyramid and take his place upon the stone bench that topped it. As they reached the step below the sacred pinnacle Ko-tan continued as though to mount to his throne, but Tarzan laid a detaining hand upon his arm. "None may sit upon a level with the gods," he admonished, stepping confidently up and seating himself upon the throne. The abashed Ko-tan showed his embarrassment, an embarrassment he feared to voice lest he incur the wrath of the king of kings. "But," added Tarzan, "a god may honor his faithful servant by inviting him to a place at his side. Come, Ko-tan; thus would I honor you in the name of Jad-ben-Otho." The ape-man's policy had for its basis an attempt not only to arouse the fearful respect of Ko-tan but to do it without making of him an enemy at heart, for he did not know how strong a hold the religion of the Ho-don had upon them, for since the time that he had prevented Ta-den and Om-at from quarreling over a religious difference the subject had been utterly taboo among them. He was therefore quick to note the evident though wordless resentment of Ko-tan at the suggestion that he entirely relinquish his throne to his guest. On the whole, however, the effect had been satisfactory as he could see from the renewed evidence of awe upon the faces of the warriors. At Tarzan's direction the business of the court continued where it had been interrupted by his advent. It consisted principally in the settling of disputes between warriors. There was present one who stood upon the step just below the throne and which Tarzan was to learn was the place reserved for the higher chiefs of the allied tribes which made up Ko-tan's kingdom. The one who attracted Tarzan's attention was a stalwart warrior of powerful physique and massive, lion-like features. He was addressing Ko-tan on a question that is as old as government and that will continue in unabated importance until man ceases to exist. It had to do with a boundary dispute with one of his neighbors. The matter itself held little or no interest for Tarzan, but he was impressed by the appearance of the speaker and when Ko-tan addressed him as Ja-don the ape-man's interest was permanently crystallized, for Ja-don was the father of Ta-den. That the knowledge would benefit him in any way seemed rather a remote possibility since he could not reveal to Ja-don his friendly relations with his son without admitting the falsity of his claims to godship. When the affairs of the audience were concluded Ko-tan suggested that the son of Jad-ben-Otho might wish to visit the temple in which were performed the religious rites coincident to the worship of the Great God. And so the ape-man was conducted by the king himself, followed by the warriors of his court, through the corridors of the palace toward the northern end of the group of buildings within the royal enclosure. The temple itself was really a part of the palace and similar in architecture. There were several ceremonial places of varying sizes, the purposes of which Tarzan could only conjecture. Each had an altar in the west end and another in the east and were oval in shape, their longest diameter lying due east and west. Each was excavated from the summit of a small hillock and all were without roofs. The western altars invariably were a single block of stone the top of which was hollowed into an oblong basin. Those at the eastern ends were similar blocks of stone with flat tops and these latter, unlike those at the opposite ends of the ovals were invariably stained or painted a reddish brown, nor did Tarzan need to examine them closely to be assured of what his keen nostrils already had told him—that the brown stains were dried and drying human blood. Below these temple courts were corridors and apartments reaching far into the bowels of the hills, dim, gloomy passages that Tarzan glimpsed as he was led from place to place on his tour of inspection of the temple. A messenger had been dispatched by Ko-tan to announce the coming visit of the son of Jad-ben-Otho with the result that they were accompanied through the temple by a considerable procession of priests whose distinguishing mark of profession seemed to consist in grotesque headdresses; sometimes hideous faces carved from wood and entirely concealing the countenances of their wearers, or again, the head of a wild beast cunningly fitted over the head of a man. The high priest alone wore no such head-dress. He was an old man with close-set, cunning eyes and a cruel, thin-lipped mouth. At first sight of him Tarzan realized that here lay the greatest danger to his ruse, for he saw at a glance that the man was antagonistic toward him and his pretensions, and he knew too that doubtless of all the people of Pal-ul-don the high priest was most likely to harbor the truest estimate of Jad-ben-Otho, and, therefore, would look with suspicion on one who claimed to be the son of a fabulous god. No matter what suspicion lurked within his crafty mind, Lu-don, the high priest of A-lur, did not openly question Tarzan's right to the title of Dor-ul-Otho, and it may be that he was restrained by the same doubts which had originally restrained Ko-tan and his warriors—the doubt that is at the bottom of the minds of all blasphemers even and which is based upon the fear that after all there may be a god. So, for the time being at least Lu-don played safe. Yet Tarzan knew as well as though the man had spoken aloud his inmost thoughts that it was in the heart of the high priest to tear the veil from his imposture. At the entrance to the temple Ko-tan had relinquished the guidance of the guest to Lu-don and now the latter led Tarzan through those portions of the temple that he wished him to see. He showed him the great room where the votive offerings were kept, gifts from the barbaric chiefs of Pal-ul-don and from their followers. These things ranged in value from presents of dried fruits to massive vessels of beaten gold, so that in the great main storeroom and its connecting chambers and corridors was an accumulation of wealth that amazed even the eyes of the owner of the secret of the treasure vaults of Opar. Moving to and fro throughout the temple were sleek black Waz-don slaves, fruits of the Ho-don raids upon the villages of their less civilized neighbors. As they passed the barred entrance to a dim corridor, Tarzan saw within a great company of pithecanthropi of all ages and of both sexes, Ho-don as well as Waz-don, the majority of them squatted upon the stone floor in attitudes of utter dejection while some paced back and forth, their features stamped with the despair of utter hopelessness. "And who are these who lie here thus unhappily?" he asked of Lu-don. It was the first question that he had put to the high priest since entering the temple, and instantly he regretted that he had asked it, for Lu-don turned upon him a face upon which the expression of suspicion was but thinly veiled. "Who should know better than the son of Jad-ben-Otho?" he retorted. "The questions of Dor-ul-Otho are not with impunity answered with other questions," said the ape-man quietly, "and it may interest Lu-don, the high priest, to know that the blood of a false priest upon the altar of his temple is not displeasing in the eyes of Jad-ben-Otho." Lu-don paled as he answered Tarzan's question. "They are the offerings whose blood must refresh the eastern altars as the sun returns to your father at the day's end." "And who told you," asked Tarzan, "that Jad-ben-Otho was pleased that his people were slain upon his altars? What if you were mistaken?" "Then countless thousands have died in vain," replied Lu-don. Ko-tan and the surrounding warriors and priests were listening attentively to the dialogue. Some of the poor victims behind the barred gateway had heard and rising, pressed close to the barrier through which one was conducted just before sunset each day, never to return. "Liberate them!" cried Tarzan with a wave of his hand toward the imprisoned victims of a cruel superstition, "for I can tell you in the name of Jad-ben-Otho that you are mistaken." About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Edgar Rice Burroughs (2020). Tarzan the Terrible. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/2020/pg2020-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at www.gutenberg.org , located at https://www.gutenberg.org/policy/license.html .

Blood-Stained Altars

Write on Interaction, Reputation, or Trust, Win up to $6000 in Prizes

ইন্টারঅ্যাকশন, রেপুটেশন বা বিশ্বাসের উপর লিখুন, পুরস্কারে $6000 পর্যন্ত জিতে নিন

Schreiben Sie über Interaktion, Reputation oder Vertrauen und gewinnen Sie Preise im Wert von bis zu 6.000 US-Dollar.

Escriba sobre interacción, reputación o confianza, gane hasta $6000 en premios

Écrivez sur l'interaction, la réputation ou la confiance, gagnez jusqu'à 6 000 $ en prix

बातचीत, प्रतिष्ठा या विश्वास पर लिखें, 6000 डॉलर तक का पुरस्कार जीतें

交流、評判、信頼について書いて、最大 6,000 ドルの賞金を獲得しましょう

상호 작용, 평판 또는 신뢰에 대해 글을 쓰고 최대 $6000의 상금을 받으세요

Escreva sobre interação, reputação ou confiança e ganhe até US$ 6.000 em prêmios

Напишите о взаимодействии, репутации или доверии, выиграйте до 6000 долларов США в виде призов

Etkileşim, İtibar veya Güven Üzerine Yazın, 6000$'a Varan Ödül Kazanın

Viết về sự tương tác, danh tiếng hoặc sự tin cậy, giành được giải thưởng lên tới $6000

撰写有关互动、声誉或信任的文章，赢取高达 6000 美元的奖金

10 LeetCode Patterns to Solve 1000 LeetCode Problems

Craig Wright's Bitcoin White Paper Document Questioned for Forgery

COPA Sues Craig Wright for Pretending to be Bitcoin founder, Satoshi Nakamoto 

The first instruction you learned should be the least you use. TL;DR: Remove all your Accidental IF-sentences Problems Addressed Code Duplication Possible Typos and defects Related Code Smells Code Smell 07 - Boolean Variables Code Smell 36 - Switch/case/elseif/else/if statements Code Smell 133 - Hardcoded IF Conditions Code Smell 156 - Implicit Else Code Smell 119 - Stairs Code Code Smell 145 - Short Circuit Hack Code Smell 101 - Comparison Against Booleans Code Smell 45 - Not Polymorphic Steps Find or Create a Polymorphic Hierarchy.


Move the Body of Each IF to the Corresponding Abstraction.


Name the Abstractions.


Name the Method.


Replace if Statements with Polymorphic Message Sends. Sample Code Before public String handleMicrophoneState(String state) {
    if (state.equals("off")) {
        return "Microphone is off";
    } else {
        return "Microphone is on";
    }
}

/* The constant representing the 'off' state is
duplicated throughout the code, 
increasing the risk of typos and spelling mistakes. 
The "else" condition doesn't explicitly check for the 'on' state;
it implicitly handles any state that is 'not off'. 
This approach leads to repetition of the IF condition
wherever the state needs handling, 
exposing internal representation and violating encapsulation.
The algorithm is not open for extension and closed for modification,
meaning that adding a new state 
will require changes in multiple places in the code. */ After // Step 1: Find or Create a Polymorphic Hierarchy

abstract class MicrophoneState { }
final class On extends MicrophoneState { }
final class Off extends MicrophoneState { }

// Step 2: Move the Body of Each IF to the Corresponding Abstraction

abstract class MicrophoneState {
    public abstract String polymorphicMethodFromIf();
}

final class On extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is on";
    }
}

final class Off extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is off";
    }
}

// Step 3: Name the Abstractions

abstract class MicrophoneState {}
final class MicrophoneStateOn extends MicrophoneState {}
final class MicrophoneStateOff extends MicrophoneState {}

// Step 4: Name the Method

abstract class MicrophoneState {
   public abstract String handle();
}

final class MicrophoneStateOn extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is on";
    }
}

final class MicrophoneStateOff extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is off";
    }
}

// Step 5: Replace if Statements with Polymorphic Message Sends

 public String handleMicrophoneState(String state) {
        Map<String, MicrophoneState> states = new HashMap<>();
        states.put("muted", new Muted());
        states.put("recording", new Recording());
        states.put("idle", new Idle());

        MicrophoneState microphoneState = 
            states.getOrDefault(state, new NullMicrophoneState());
        return microphoneState.handle();
    } Type [x]Semi-Automatic Safety Most steps are mechanic. This is a pretty safe refactoring. Why is the code better? The refactored code follows the open/closed principle and favors polymorphism instead of using IFs. Limitations You should only apply it to Accidental IFs. Leave the business rules as "domain ifs," and don't apply this refactoring. Tags IFs Related Refactorings https://hackernoon.com/refactoring-013-eliminating-repeated-code-with-dry-principles?embedable=true See Also https://hackernoon.com/how-to-get-rid-of-annoying-ifs-forever-zuh3zlo?embedable=true This article is part of the Refactoring Series. How to Improve your Code With Easy Refactorings The first instruction you learned should be the least you use. The first instruction you learned should be the least you use. TL;DR: Remove all your Accidental IF-sentences TL;DR: Remove all your Accidental IF-sentences Problems Addressed Code Duplication Code Duplication Possible Typos and defects Possible Typos and defects Related Code Smells Code Smell 07 - Boolean Variables Code Smell 07 - Boolean Variables Code Smell 36 - Switch/case/elseif/else/if statements Code Smell 36 - Switch/case/elseif/else/if statements Code Smell 133 - Hardcoded IF Conditions Code Smell 133 - Hardcoded IF Conditions Code Smell 156 - Implicit Else Code Smell 156 - Implicit Else Code Smell 119 - Stairs Code Code Smell 119 - Stairs Code Code Smell 145 - Short Circuit Hack Code Smell 145 - Short Circuit Hack Code Smell 101 - Comparison Against Booleans Code Smell 101 - Comparison Against Booleans Code Smell 45 - Not Polymorphic Code Smell 45 - Not Polymorphic Steps Find or Create a Polymorphic Hierarchy. Move the Body of Each IF to the Corresponding Abstraction. Name the Abstractions. Name the Method. Replace if Statements with Polymorphic Message Sends. Find or Create a Polymorphic Hierarchy. Find or Create a Polymorphic Hierarchy. Move the Body of Each IF to the Corresponding Abstraction. Move the Body of Each IF to the Corresponding Abstraction. Name the Abstractions. Name the Abstractions. Name the Method. Name the Method. Replace if Statements with Polymorphic Message Sends. Replace if Statements with Polymorphic Message Sends. Sample Code Before public String handleMicrophoneState(String state) {
    if (state.equals("off")) {
        return "Microphone is off";
    } else {
        return "Microphone is on";
    }
}

/* The constant representing the 'off' state is
duplicated throughout the code, 
increasing the risk of typos and spelling mistakes. 
The "else" condition doesn't explicitly check for the 'on' state;
it implicitly handles any state that is 'not off'. 
This approach leads to repetition of the IF condition
wherever the state needs handling, 
exposing internal representation and violating encapsulation.
The algorithm is not open for extension and closed for modification,
meaning that adding a new state 
will require changes in multiple places in the code. */ public String handleMicrophoneState(String state) {
    if (state.equals("off")) {
        return "Microphone is off";
    } else {
        return "Microphone is on";
    }
}

/* The constant representing the 'off' state is
duplicated throughout the code, 
increasing the risk of typos and spelling mistakes. 
The "else" condition doesn't explicitly check for the 'on' state;
it implicitly handles any state that is 'not off'. 
This approach leads to repetition of the IF condition
wherever the state needs handling, 
exposing internal representation and violating encapsulation.
The algorithm is not open for extension and closed for modification,
meaning that adding a new state 
will require changes in multiple places in the code. */ After // Step 1: Find or Create a Polymorphic Hierarchy

abstract class MicrophoneState { }
final class On extends MicrophoneState { }
final class Off extends MicrophoneState { }

// Step 2: Move the Body of Each IF to the Corresponding Abstraction

abstract class MicrophoneState {
    public abstract String polymorphicMethodFromIf();
}

final class On extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is on";
    }
}

final class Off extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is off";
    }
}

// Step 3: Name the Abstractions

abstract class MicrophoneState {}
final class MicrophoneStateOn extends MicrophoneState {}
final class MicrophoneStateOff extends MicrophoneState {}

// Step 4: Name the Method

abstract class MicrophoneState {
   public abstract String handle();
}

final class MicrophoneStateOn extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is on";
    }
}

final class MicrophoneStateOff extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is off";
    }
}

// Step 5: Replace if Statements with Polymorphic Message Sends

 public String handleMicrophoneState(String state) {
        Map<String, MicrophoneState> states = new HashMap<>();
        states.put("muted", new Muted());
        states.put("recording", new Recording());
        states.put("idle", new Idle());

        MicrophoneState microphoneState = 
            states.getOrDefault(state, new NullMicrophoneState());
        return microphoneState.handle();
    } // Step 1: Find or Create a Polymorphic Hierarchy

abstract class MicrophoneState { }
final class On extends MicrophoneState { }
final class Off extends MicrophoneState { }

// Step 2: Move the Body of Each IF to the Corresponding Abstraction

abstract class MicrophoneState {
    public abstract String polymorphicMethodFromIf();
}

final class On extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is on";
    }
}

final class Off extends MicrophoneState {
    @Override
    public String polymorphicMethodFromIf() {
        return "Microphone is off";
    }
}

// Step 3: Name the Abstractions

abstract class MicrophoneState {}
final class MicrophoneStateOn extends MicrophoneState {}
final class MicrophoneStateOff extends MicrophoneState {}

// Step 4: Name the Method

abstract class MicrophoneState {
   public abstract String handle();
}

final class MicrophoneStateOn extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is on";
    }
}

final class MicrophoneStateOff extends MicrophoneState {
    @Override
    String handle() {
        return "Microphone is off";
    }
}

// Step 5: Replace if Statements with Polymorphic Message Sends

 public String handleMicrophoneState(String state) {
        Map<String, MicrophoneState> states = new HashMap<>();
        states.put("muted", new Muted());
        states.put("recording", new Recording());
        states.put("idle", new Idle());

        MicrophoneState microphoneState = 
            states.getOrDefault(state, new NullMicrophoneState());
        return microphoneState.handle();
    } Type [x]Semi-Automatic [x] Semi-Automatic [x] Safety Most steps are mechanic. This is a pretty safe refactoring. Why is the code better? The refactored code follows the open/closed principle and favors polymorphism instead of using IFs. Limitations You should only apply it to Accidental IFs . Accidental IFs Accidental IFs Leave the business rules as "domain ifs," and don't apply this refactoring. "domain ifs," Tags IFs IFs Related Refactorings https://hackernoon.com/refactoring-013-eliminating-repeated-code-with-dry-principles?embedable=true https://hackernoon.com/refactoring-013-eliminating-repeated-code-with-dry-principles?embedable=true See Also https://hackernoon.com/how-to-get-rid-of-annoying-ifs-forever-zuh3zlo?embedable=true https://hackernoon.com/how-to-get-rid-of-annoying-ifs-forever-zuh3zlo?embedable=true This article is part of the Refactoring Series. How to Improve your Code With Easy Refactorings How to Improve your Code With Easy Refactorings

Refactoring 014 - How to Remove IF

Authors:
(1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA;
(2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA;
(3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA;
(4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy;
(5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA;
(6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA;
(7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Table of Links Abstract and I Introduction II Methods III Results IV Discussion V. Conclusion, Acknowledgment, and References V. CONCLUSION In this study, we demonstrated the feasibility of employing deep learning in gait state detection for exoskeleton control, eliminating the necessity of ground reaction force sensors. Despite challenges such as a limited dataset, realtime constraints, and issues related to error propagation, these obstacles were successfully addressed. Evaluation of our model’s closed-loop performance versus typical sensors (i.e., force plates, FSR footplates) highlights the system’s applicability across new users and several walking speeds and conditions. To address our model limitations, future work could include training datasets encompassing various walking speeds, overground walking and other activities (e.g. ramps, stairs). Furthermore, this work could be used to implement a state machine that utilizes an estimated stance interpolation factor, selecting impedance parameters in the mid-level controller for exoskeleton control and tailored assistance during walking. The proposed framework will be validated in future on additional healthy individuals as well as individuals with lower-limb impairments (e.g., stroke, spinal cord injury). ACKNOWLEDGMENT This work was supported by the National Science Foundation / National Robotics Initiative (Grant No: 2024488). We would like to thank Tim Haswell for his technical support on the hardware improvements of the ExoMotus-X2 exoskeleton. REFERENCES [1] R. Baud, A. R. Manzoori, A. Ijspeert, and M. Bouri, “Review of control strategies for lower-limb exoskeletons to assist gait,” Journal of NeuroEngineering and Rehabilitation, vol. 18, no. 1, Jul. 2021. [Online]. Available: https://doi.org/10.1186/s12984-021-00906-3 [2] J. de Miguel-Fernandez, J. Lobo-Prat, E. Prinsen, J. M. Font-Llagunes, and L. Marchal-Crespo, “Control strategies used in lower limb exoskeletons for gait rehabilitation after brain injury: a systematic review and analysis of clinical effectiveness,” Journal of NeuroEngineering and Rehabilitation, vol. 20, no. 1, Feb. 2023. [3] E. Basalp, P. Wolf, and L. Marchal-Crespo, “Haptic training: Which types facilitate (re)learning of which motor task and for whom? answers by a review,” IEEE Transactions on Haptics, vol. 14, no. 4, pp. 722–739, 2021. [4] M. Kim, A. M. Simon, and L. J. Hargrove, “Seamless and intuitive control of a powered prosthetic leg using deep neural network for transfemoral amputees,” Wearable Technologies, vol. 3, 2022. [5] W. Huo, S. Mohammed, Y. Amirat, and K. Kong, “Fast gait mode detection and assistive torque control of an exoskeletal robotic orthosis for walking assistance,” IEEE Transactions on Robotics, vol. 34, no. 4, pp. 1035–1052, 2018. [6] A. Ortlieb, R. Baud, T. Tracchia, B. Denkinger, Q. Herzig, H. Bleuler, and M. Bouri, “An active impedance controller to assist gait in people with neuromuscular diseases: Implementation to the hip joint of the autonomyo exoskeleton,” in 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), 2018. [7] E. B. Kuc¸¨ uktabak, Y. Wen, S. J. Kim, M. R. Short, D. Ludvig, L. Hargrove, E. J. Perreault, K. M. Lynch, and J. L. Pons, “Haptic transparency and interaction force control for a lower-limb exoskeleton,” IEEE Transactions on Robotics, pp. 1–19, 2024. [8] M. Mundt, A. Koeppe, S. David, F. Bamer, W. Potthast, and B. Markert, “Prediction of ground reaction force and joint moments based on optical motion capture data during gait,” Medical Engineering & Physics, vol. 86, pp. 29–34, 2020. [9] A. Karatsidis, G. Bellusci, H. Schepers, M. de Zee, M. Andersen, and P. Veltink, “Estimation of ground reaction forces and moments during gait using only inertial motion capture,” Sensors, vol. 17, no. 12, p. 75, Dec. 2016. [Online]. Available: https://doi.org/10.3390/s17010075 [10] L. Mourot, L. Hoyet, F. L. Clerc, and P. Hellier, “Underpressure: Deep learning for foot contact detection, ground reaction force estimation and footskate cleanup,” Computer Graphics Forum, vol. 41, no. 8, pp. 195–206, 2022. [11] J.-Y. Jung, W. Heo, H. Yang, and H. Park, “A neural network-based gait phase classification method using sensors equipped on lower limb exoskeleton robots,” Sensors, vol. 15, no. 11, pp. 27 738–27 759, 2015. [12] D.-X. Liu, X. Wu, W. Du, C. Wang, and T. Xu, “Gait phase recognition for lower-limb exoskeleton with only joint angular sensors,” Sensors, vol. 16, no. 10, p. 1579, Sep. 2016. [13] C. Camardella, F. Porcini, A. Filippeschi, S. Marcheschi, M. Solazzi, and A. Frisoli, “Gait phases blended control for enhancing transparency on lower-limb exoskeletons,” IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5453–5460, 2021. [14] V. Lippi, C. Camardella, A. Filippeschi, and F. Porcini, “Identification of gait phases with neural networks for smooth transparent control of a lower limb exoskeleton,” in Proceedings of the 18th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,, INSTICC. SciTePress, 2021, pp. 171–178. [15] J. Fong, E. B. Kuc¸uktabak, V. Crocher, Y. Tan, K. M. Lynch, J. L. Pons, and D. Oetomo, “Canopen robot controller (corc): An open software stack for human robot interaction development,” in Wearable Robotics: Challenges and Trends, J. C. Moreno, J. Masood, U. Schneider, C. Maufroy, and J. L. Pons, Eds. Cham: Springer International Publishing, 2022, pp. 287–292. [16] F. Labarriere, E. Thomas, L. Calistri, V. Optasanu, M. Gueugnon, ` P. Ornetti, and D. Laroche, “Machine learning approaches for activity recognition and/or activity prediction in locomotion assistive devices—a systematic review,” Sensors, vol. 20, no. 21, 2020. [17] M. Kim and L. J. Hargrove, “Deep-learning to map a benchmark dataset of non-amputee ambulation for controlling an open source bionic leg,” IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 10 597–10 604, 2022. [18] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014. [19] E. B. Kuc¸uktabak, Y. Wen, M. Short, E. Demirbas¸, K. Lynch, and J. Pons, “Virtual physical coupling of two lower-limb exoskeletons,” in 2023 International Conference on Rehabilitation Robotics (ICORR), 2023, pp. 1–6. [20] S. Levine, A. Kumar, G. Tucker, and J. Fu, “Offline reinforcement learning: Tutorial, review, and perspectives on open problems,” arXiv preprint arXiv:2005.01643, 2020. [21] R. A. Alamro, A. E. Chisholm, A. M. M. Williams, M. G. Carpenter, and T. Lam, “Overground walking with a robotic exoskeleton elicits trunk muscle activity in people with high-thoracic motor-complete spinal cord injury,” Journal of NeuroEngineering and Rehabilitation, vol. 15, no. 1, Nov. 2018. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy; (5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA; (6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Authors: Authors: (1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy; (5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA; (6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Table of Links Abstract and I Introduction Abstract and I Introduction II Methods II Methods III Results III Results IV Discussion IV Discussion V. Conclusion, Acknowledgment, and References V. Conclusion, Acknowledgment, and References V. CONCLUSION In this study, we demonstrated the feasibility of employing deep learning in gait state detection for exoskeleton control, eliminating the necessity of ground reaction force sensors. Despite challenges such as a limited dataset, realtime constraints, and issues related to error propagation, these obstacles were successfully addressed. Evaluation of our model’s closed-loop performance versus typical sensors (i.e., force plates, FSR footplates) highlights the system’s applicability across new users and several walking speeds and conditions. To address our model limitations, future work could include training datasets encompassing various walking speeds, overground walking and other activities (e.g. ramps, stairs). Furthermore, this work could be used to implement a state machine that utilizes an estimated stance interpolation factor, selecting impedance parameters in the mid-level controller for exoskeleton control and tailored assistance during walking. The proposed framework will be validated in future on additional healthy individuals as well as individuals with lower-limb impairments (e.g., stroke, spinal cord injury). ACKNOWLEDGMENT This work was supported by the National Science Foundation / National Robotics Initiative (Grant No: 2024488). We would like to thank Tim Haswell for his technical support on the hardware improvements of the ExoMotus-X2 exoskeleton. REFERENCES [1] R. Baud, A. R. Manzoori, A. Ijspeert, and M. Bouri, “Review of control strategies for lower-limb exoskeletons to assist gait,” Journal of NeuroEngineering and Rehabilitation, vol. 18, no. 1, Jul. 2021. [Online]. Available: https://doi.org/10.1186/s12984-021-00906-3 [2] J. de Miguel-Fernandez, J. Lobo-Prat, E. Prinsen, J. M. Font-Llagunes, and L. Marchal-Crespo, “Control strategies used in lower limb exoskeletons for gait rehabilitation after brain injury: a systematic review and analysis of clinical effectiveness,” Journal of NeuroEngineering and Rehabilitation, vol. 20, no. 1, Feb. 2023. [3] E. Basalp, P. Wolf, and L. Marchal-Crespo, “Haptic training: Which types facilitate (re)learning of which motor task and for whom? answers by a review,” IEEE Transactions on Haptics, vol. 14, no. 4, pp. 722–739, 2021. [4] M. Kim, A. M. Simon, and L. J. Hargrove, “Seamless and intuitive control of a powered prosthetic leg using deep neural network for transfemoral amputees,” Wearable Technologies, vol. 3, 2022. [5] W. Huo, S. Mohammed, Y. Amirat, and K. Kong, “Fast gait mode detection and assistive torque control of an exoskeletal robotic orthosis for walking assistance,” IEEE Transactions on Robotics, vol. 34, no. 4, pp. 1035–1052, 2018. [6] A. Ortlieb, R. Baud, T. Tracchia, B. Denkinger, Q. Herzig, H. Bleuler, and M. Bouri, “An active impedance controller to assist gait in people with neuromuscular diseases: Implementation to the hip joint of the autonomyo exoskeleton,” in 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), 2018. [7] E. B. Kuc¸¨ uktabak, Y. Wen, S. J. Kim, M. R. Short, D. Ludvig, L. Hargrove, E. J. Perreault, K. M. Lynch, and J. L. Pons, “Haptic transparency and interaction force control for a lower-limb exoskeleton,” IEEE Transactions on Robotics, pp. 1–19, 2024. [8] M. Mundt, A. Koeppe, S. David, F. Bamer, W. Potthast, and B. Markert, “Prediction of ground reaction force and joint moments based on optical motion capture data during gait,” Medical Engineering & Physics, vol. 86, pp. 29–34, 2020. [9] A. Karatsidis, G. Bellusci, H. Schepers, M. de Zee, M. Andersen, and P. Veltink, “Estimation of ground reaction forces and moments during gait using only inertial motion capture,” Sensors, vol. 17, no. 12, p. 75, Dec. 2016. [Online]. Available: https://doi.org/10.3390/s17010075 [10] L. Mourot, L. Hoyet, F. L. Clerc, and P. Hellier, “Underpressure: Deep learning for foot contact detection, ground reaction force estimation and footskate cleanup,” Computer Graphics Forum, vol. 41, no. 8, pp. 195–206, 2022. [11] J.-Y. Jung, W. Heo, H. Yang, and H. Park, “A neural network-based gait phase classification method using sensors equipped on lower limb exoskeleton robots,” Sensors, vol. 15, no. 11, pp. 27 738–27 759, 2015. [12] D.-X. Liu, X. Wu, W. Du, C. Wang, and T. Xu, “Gait phase recognition for lower-limb exoskeleton with only joint angular sensors,” Sensors, vol. 16, no. 10, p. 1579, Sep. 2016. [13] C. Camardella, F. Porcini, A. Filippeschi, S. Marcheschi, M. Solazzi, and A. Frisoli, “Gait phases blended control for enhancing transparency on lower-limb exoskeletons,” IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5453–5460, 2021. [14] V. Lippi, C. Camardella, A. Filippeschi, and F. Porcini, “Identification of gait phases with neural networks for smooth transparent control of a lower limb exoskeleton,” in Proceedings of the 18th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,, INSTICC. SciTePress, 2021, pp. 171–178. [15] J. Fong, E. B. Kuc¸uktabak, V. Crocher, Y. Tan, K. M. Lynch, J. L. Pons, and D. Oetomo, “Canopen robot controller (corc): An open software stack for human robot interaction development,” in Wearable Robotics: Challenges and Trends, J. C. Moreno, J. Masood, U. Schneider, C. Maufroy, and J. L. Pons, Eds. Cham: Springer International Publishing, 2022, pp. 287–292. [16] F. Labarriere, E. Thomas, L. Calistri, V. Optasanu, M. Gueugnon, ` P. Ornetti, and D. Laroche, “Machine learning approaches for activity recognition and/or activity prediction in locomotion assistive devices—a systematic review,” Sensors, vol. 20, no. 21, 2020. [17] M. Kim and L. J. Hargrove, “Deep-learning to map a benchmark dataset of non-amputee ambulation for controlling an open source bionic leg,” IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 10 597–10 604, 2022. [18] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014. [19] E. B. Kuc¸uktabak, Y. Wen, M. Short, E. Demirbas¸, K. Lynch, and J. Pons, “Virtual physical coupling of two lower-limb exoskeletons,” in 2023 International Conference on Rehabilitation Robotics (ICORR), 2023, pp. 1–6. [20] S. Levine, A. Kumar, G. Tucker, and J. Fu, “Offline reinforcement learning: Tutorial, review, and perspectives on open problems,” arXiv preprint arXiv:2005.01643, 2020. [21] R. A. Alamro, A. E. Chisholm, A. M. M. Williams, M. G. Carpenter, and T. Lam, “Overground walking with a robotic exoskeleton elicits trunk muscle activity in people with high-thoracic motor-complete spinal cord injury,” Journal of NeuroEngineering and Rehabilitation, vol. 15, no. 1, Nov. 2018. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Weight Distribution Estimation in Lower-Limb Exoskeletons via Deep Learning: Conclusion

Authors:
(1) Ronen Eldan, Microsoft Research (email: roneneldan@microsoft.com);
(2) Mark Russinovich, Microsoft Azure and Both authors contributed equally to this work, (email: mark.russinovich@microsoft.com). Table of Links Abstract and Introduction
Description of our technique
Evaluation methodology
Results
Conclusion, Acknowledgment, and References
Appendix 3 Evaluation methodology To adequately assess the efficacy of our unlearning technique, our evaluation framework is grounded on two primary dimensions: preservation of general model capabilities and eradication of specific, targeted knowledge. 3.1 Preservation of General Capabilities To ensure that our method did not impair the model’s overall capabilities when prompts are unrelated to the unlearned topic, we leverage widely-accepted benchmarks like WinoGrande, HellaSwag, and piqa to objectively gauge the model’s performance and ascertain that the overarching linguistic understanding and a wide array of other capabilities remain intact. 3.2 Eradication of Targeted Knowledge The crux of our evaluation lies in determining the extent to which the model retains or has lost knowledge of the unlearned content. This evaluation component primarily involves a series of black-box tests, utilizing prompts specifically curated to elicit knowledge about the unlearned content (specifically, the Harry Potter universe), both directly and indirectly. 3.2.1 Completion-based evaluation We have curated a list of prompts in a manner that either: • Provides partial information related to the Harry Potter universe, demanding the model to complete the information based on its internal knowledge. • Offers instructions that, either overtly or covertly, might prompt the baseline model to disclose familiarity with the books. Examples of such prompts include scenarios like: ”When Harry returned to class, he observed his best friends,”, ”Draft a brief narrative in the style of Harry Potter. Short story:”, ”Narrate a tale about a boy who resides in a cupboard beneath the stairs in his relatives’ home, who are mistreating him, only to later discover he possesses magical abilities.” Prompts also delved into subtler references such as: ”While lounging beside the fireplace, the elder pupils recounted the distinct attributes of the four Hogwarts factions, describing them as” and ”Throughout the ages, numerous Defense Against the Dark Arts educators graced Hogwarts, each bearing their unique history. Pupils frequently reminisced about”. The full list can be found in the supplementary material. To ensure a comprehensive evaluation, we compiled a list of 300 such prompts with the aid of GPT-4 (included in the complementary material). GPT-4’s role was further leveraged to analyze the completions during parameter search, but due to its apparent inaccuracy at the task, for our final training, a manual inspection was conducted on the completions in the sake of additional scrutiny. 3.2.2 Token-probability-based evaluation A complementary approach for evaluation is based on inspecting completion probabilities for select prompts. For instance, for the cue ”Harry Potter studies ”, we verify that the model does not allocate high probabilities to Harry Potter-specific terms such as ”magic” or ”wizardry”. We collected a list of 30 such prompts, and (manually) categorized the possible next tokens as either content-specific or generic (further details are given in Appendix 6.2) 3.3 Open Evaluation Recognizing the intrinsic limitations of automated benchmarks and internal evaluations, we believe that unlearning verification parallels endeavors like jailbreaking in adversarial nature. Therefore, we open-sourced the model4 , encouraging the broader community to challenge it, providing a more diverse and extensive set of tests to discern if any remnants of the targeted knowledge persist. This paper is available on arxiv under CC 4.0 license. Authors: (1) Ronen Eldan, Microsoft Research (email: roneneldan@microsoft.com); (2) Mark Russinovich, Microsoft Azure and Both authors contributed equally to this work, (email: mark.russinovich@microsoft.com). Authors: Authors: (1) Ronen Eldan, Microsoft Research (email: roneneldan@microsoft.com); (2) Mark Russinovich, Microsoft Azure and Both authors contributed equally to this work, (email: mark.russinovich@microsoft.com). Table of Links Abstract and Introduction Description of our technique Evaluation methodology Results Conclusion, Acknowledgment, and References Appendix Abstract and Introduction Abstract and Introduction Description of our technique Description of our technique Evaluation methodology Evaluation methodology Results Results Conclusion, Acknowledgment, and References Conclusion, Acknowledgment, and References Appendix Appendix 3 Evaluation methodology To adequately assess the efficacy of our unlearning technique, our evaluation framework is grounded on two primary dimensions: preservation of general model capabilities and eradication of specific, targeted knowledge. 3.1 Preservation of General Capabilities To ensure that our method did not impair the model’s overall capabilities when prompts are unrelated to the unlearned topic, we leverage widely-accepted benchmarks like WinoGrande, HellaSwag, and piqa to objectively gauge the model’s performance and ascertain that the overarching linguistic understanding and a wide array of other capabilities remain intact. 3.2 Eradication of Targeted Knowledge The crux of our evaluation lies in determining the extent to which the model retains or has lost knowledge of the unlearned content. This evaluation component primarily involves a series of black-box tests, utilizing prompts specifically curated to elicit knowledge about the unlearned content (specifically, the Harry Potter universe), both directly and indirectly. 3.2.1 Completion-based evaluation 3.2.1 Completion-based evaluation We have curated a list of prompts in a manner that either: • Provides partial information related to the Harry Potter universe, demanding the model to complete the information based on its internal knowledge. • Offers instructions that, either overtly or covertly, might prompt the baseline model to disclose familiarity with the books. Examples of such prompts include scenarios like: ”When Harry returned to class, he observed his best friends,”, ”Draft a brief narrative in the style of Harry Potter. Short story:”, ”Narrate a tale about a boy who resides in a cupboard beneath the stairs in his relatives’ home, who are mistreating him, only to later discover he possesses magical abilities.” Prompts also delved into subtler references such as: ”While lounging beside the fireplace, the elder pupils recounted the distinct attributes of the four Hogwarts factions, describing them as” and ”Throughout the ages, numerous Defense Against the Dark Arts educators graced Hogwarts, each bearing their unique history. Pupils frequently reminisced about”. The full list can be found in the supplementary material. To ensure a comprehensive evaluation, we compiled a list of 300 such prompts with the aid of GPT-4 (included in the complementary material). GPT-4’s role was further leveraged to analyze the completions during parameter search, but due to its apparent inaccuracy at the task, for our final training, a manual inspection was conducted on the completions in the sake of additional scrutiny. 3.2.2 Token-probability-based evaluation 3.2.2 Token-probability-based evaluation A complementary approach for evaluation is based on inspecting completion probabilities for select prompts. For instance, for the cue ”Harry Potter studies ”, we verify that the model does not allocate high probabilities to Harry Potter-specific terms such as ”magic” or ”wizardry”. We collected a list of 30 such prompts, and (manually) categorized the possible next tokens as either content-specific or generic (further details are given in Appendix 6.2) 3.3 Open Evaluation Recognizing the intrinsic limitations of automated benchmarks and internal evaluations, we believe that unlearning verification parallels endeavors like jailbreaking in adversarial nature. Therefore, we open-sourced the model4 , encouraging the broader community to challenge it, providing a more diverse and extensive set of tests to discern if any remnants of the targeted knowledge persist. This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

Who’s Harry Potter? Approximate Unlearning in LLMs:  Evaluation methodology

Authors:
(1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA;
(2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA;
(3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA;
(4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy;
(5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA;
(6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA;
(7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Table of Links Abstract and I Introduction II Methods III Results IV Discussion V. Conclusion, Acknowledgment, and References Abstract In the control of lower-limb exoskeletons with feet, the phase in the gait cycle can be identified by monitoring the weight distribution at the feet. This phase information can be used in the exoskeleton’s controller to compensate the dynamics of the exoskeleton and to assign impedance parameters. Typically the weight distribution is calculated using data from sensors such as treadmill force plates or insole force sensors. However, these solutions increase both the setup complexity and cost. For this reason, we propose a deep-learning approach that uses a short time window of joint kinematics to predict the weight distribution of an exoskeleton in real time. The model was trained on treadmill walking data from six users wearing a four-degree-of-freedom exoskeleton and tested in real time on three different users wearing the same device. This test set includes two users not present in the training set to demonstrate the model’s ability to generalize across individuals. Results show that the proposed method is able to fit the actual weight distribution with R 2 = 0.9 and is suitable for real-time control with prediction times less than 1 ms. Experiments in closed-loop exoskeleton control show that deep-learning-based weight distribution estimation can be used to replace force sensors in overground and treadmill walking. I. INTRODUCTION Lower-limb exoskeletons are increasingly being utilized in everyday tasks and rehabilitation for individuals with gait impairments. In clinical settings, these exoskeletons are typically classified into two main types: partial assistance and full mobilization. Full mobilization exoskeletons are designed for patients with severe motor control disorders, providing autonomous movement of the legs regardless of the patient’s input [1]. Conversely, partial assistance exoskeletons aid in the patient’s movement while still requiring their active input. In this paper, we focus on partial assistance exoskeletons, as they are commonly employed in rehabilitation. Depending on the functional level of the patient, these exoskeletons use techniques like haptic guidance/assistance [2] or error augmentation/resistance [3] to facilitate the (re)learning of walking behaviors. Implementing both haptic assistance and resistance strategies requires precise control over the interaction torques between the user and the exoskeleton, ensuring effective and safe physical interaction. Typically, the control of the interaction between a user and exoskeleton adopts a hierarchy of control levels: high-level, mid-level, and low-level [1], [4]. For lower-limb exoskeletons, the high-level controller calculates the desired interaction torques tailored to specific ambulatory activities, such as walking on flat terrain, stairs or ramps. The role of the mid-level controller is to estimate the various states within an activity, for example, identifying the gait states (e.g., swing and stance) during walking and setting a desired interaction torque behavior accordingly. Finally, the low-level controller is responsible for compensating the exoskeleton’s dynamics and generating motor commands, based on the desired interaction torque profile identified by the mid-level controller. In this three-level approach, an accurate, real-time estimation of the gait state in the mid-level and low-level is essential for properly controlling the exoskeleton. A common approach to gait state detection is to use gait events (e.g., heel strike, toe-off) to identify the switching between discrete gait states: left stance, right stance, and double stance [5], [6], [1]. However, achieving a smooth transition between these states is vital to ensure a control input that limits unexpected behaviors for the user. Kuc¸¨ uktabak et al. [7] proposed a ¨ whole-exoskeleton closed-loop compensation (WECC) controller that uses the ratio of ground reaction forces (GRF) (i.e., weight distribution) to approximate the double stance phase as a transition between the left and right stance. The weight distribution, quantified as the stance interpolation factor α, smoothly changes from zero (left stance) to one (right stance) during double stance, as visualized in Fig. 1. Nonetheless, the requirement of additional sensors for GRF measurements, such as treadmill force plates or insoles, increases the complexity and cost of these systems. Machine learning models have demonstrated their ability to estimate gait states using only joint kinematics of humans walking without exoskeletons, thus eliminating the need for GRF measurements [8], [9], [10]. Motivated by these results, recent studies have proposed real-time estimation of gait states with lower-limb exoskeletons. For instance, Jung et al. [11] and Liu et al. [12] used data from joint encoders and an inertial measurement unit (IMU) positioned on the trunk to classify the gait cycle during walking. These studies divided gait cycles into two (i.e., stance and swing) and eight gait states, respectively. However, discrete control states result in torque or modeling discontinuities. Camardella et al. [13] and Lippi et al. [14] address this issue by implementing a linear regression and a Neural Network (NN) model for regression with joint angles as inputs. The model’s output smoothly transitions between left and right stance, providing a continuous representation of the gait state. Despite this advancement, these approaches present some limitations. First, they only consider the present joint configuration, discarding past information. This approach might lead to inaccuracies in predictions as the weight distribution is influenced not only by the current posture, but also by the velocity and acceleration of the motion. Moreover, few considerations in these studies are given to the generalization across subjects and the prediction time performances. The latter is essential for real-time implementation of the exoskeleton’s whole-body controller. To address the aforementioned limitations, we present a deep-learning approach for estimating the weight distribution of a user wearing a lower-limb exoskeleton. This approach utilizes a short time window of kinematic data extracted from sensors integrated in the exoskeleton. For the validation of predictions, we compared the proposed approach with the method in [14] which only uses the instantaneous posture. Employing distinct users for training and testing resulted in a user-independent estimation, eliminating the need for a calibration phase. The proposed deep-learning method was trained on walking data from six users, and its closed-loop performance was evaluated on three users (two of which were not included in the training set) during treadmill and overground walking. The results showed that weight distribution can be accurately estimated using kinematic data with an R 2 of 0.9 averaged across users. In addition, this approach was compatible with real-time requirements, achieving a prediction time less than 1 ms. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy; (5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA; (6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Authors: Authors: (1) Clement Lhos, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (2) Emek Barıs¸ Kuc¸uktabak, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (3) Lorenzo Vianello, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA; (4) Lorenzo Amato, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and The Biorobotics Institute, Scuola Superiore Sant’Anna, 56025 Pontedera, Italy and Department of Excellence in Robotics & AI, Scuola Superiore Sant’Anna, 56127 Pisa, Italy; (5) Matthew R. Short, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA; (6) Kevin Lynch2, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA; (7) Jose L. Pons, Legs and Walking Lab of Shirley Ryan AbilityLab, Chicago, IL, USA, Center for Robotics and Biosystems, Northwestern University, Evanston, IL, USA and Department of Biomedical Engineering, Northwestern University, Evanston, IL, USA. Table of Links Abstract and I Introduction Abstract and I Introduction II Methods II Methods III Results III Results IV Discussion IV Discussion V. Conclusion, Acknowledgment, and References V. Conclusion, Acknowledgment, and References Abstract In the control of lower-limb exoskeletons with feet, the phase in the gait cycle can be identified by monitoring the weight distribution at the feet. This phase information can be used in the exoskeleton’s controller to compensate the dynamics of the exoskeleton and to assign impedance parameters. Typically the weight distribution is calculated using data from sensors such as treadmill force plates or insole force sensors. However, these solutions increase both the setup complexity and cost. For this reason, we propose a deep-learning approach that uses a short time window of joint kinematics to predict the weight distribution of an exoskeleton in real time. The model was trained on treadmill walking data from six users wearing a four-degree-of-freedom exoskeleton and tested in real time on three different users wearing the same device. This test set includes two users not present in the training set to demonstrate the model’s ability to generalize across individuals. Results show that the proposed method is able to fit the actual weight distribution with R 2 = 0.9 and is suitable for real-time control with prediction times less than 1 ms. Experiments in closed-loop exoskeleton control show that deep-learning-based weight distribution estimation can be used to replace force sensors in overground and treadmill walking. I. INTRODUCTION Lower-limb exoskeletons are increasingly being utilized in everyday tasks and rehabilitation for individuals with gait impairments. In clinical settings, these exoskeletons are typically classified into two main types: partial assistance and full mobilization. Full mobilization exoskeletons are designed for patients with severe motor control disorders, providing autonomous movement of the legs regardless of the patient’s input [1]. Conversely, partial assistance exoskeletons aid in the patient’s movement while still requiring their active input. In this paper, we focus on partial assistance exoskeletons, as they are commonly employed in rehabilitation. Depending on the functional level of the patient, these exoskeletons use techniques like haptic guidance/assistance [2] or error augmentation/resistance [3] to facilitate the (re)learning of walking behaviors. Implementing both haptic assistance and resistance strategies requires precise control over the interaction torques between the user and the exoskeleton, ensuring effective and safe physical interaction. Typically, the control of the interaction between a user and exoskeleton adopts a hierarchy of control levels: high-level, mid-level, and low-level [1], [4]. For lower-limb exoskeletons, the high-level controller calculates the desired interaction torques tailored to specific ambulatory activities, such as walking on flat terrain, stairs or ramps. The role of the mid-level controller is to estimate the various states within an activity, for example, identifying the gait states (e.g., swing and stance) during walking and setting a desired interaction torque behavior accordingly. Finally, the low-level controller is responsible for compensating the exoskeleton’s dynamics and generating motor commands, based on the desired interaction torque profile identified by the mid-level controller. In this three-level approach, an accurate, real-time estimation of the gait state in the mid-level and low-level is essential for properly controlling the exoskeleton. A common approach to gait state detection is to use gait events (e.g., heel strike, toe-off) to identify the switching between discrete gait states: left stance, right stance, and double stance [5], [6], [1]. However, achieving a smooth transition between these states is vital to ensure a control input that limits unexpected behaviors for the user. Kuc¸¨ uktabak et al. [7] proposed a ¨ whole-exoskeleton closed-loop compensation (WECC) controller that uses the ratio of ground reaction forces (GRF) (i.e., weight distribution) to approximate the double stance phase as a transition between the left and right stance. The weight distribution, quantified as the stance interpolation factor α, smoothly changes from zero (left stance) to one (right stance) during double stance, as visualized in Fig. 1. Nonetheless, the requirement of additional sensors for GRF measurements, such as treadmill force plates or insoles, increases the complexity and cost of these systems. Machine learning models have demonstrated their ability to estimate gait states using only joint kinematics of humans walking without exoskeletons, thus eliminating the need for GRF measurements [8], [9], [10]. Motivated by these results, recent studies have proposed real-time estimation of gait states with lower-limb exoskeletons. For instance, Jung et al. [11] and Liu et al. [12] used data from joint encoders and an inertial measurement unit (IMU) positioned on the trunk to classify the gait cycle during walking. These studies divided gait cycles into two (i.e., stance and swing) and eight gait states, respectively. However, discrete control states result in torque or modeling discontinuities. Camardella et al. [13] and Lippi et al. [14] address this issue by implementing a linear regression and a Neural Network (NN) model for regression with joint angles as inputs. The model’s output smoothly transitions between left and right stance, providing a continuous representation of the gait state. Despite this advancement, these approaches present some limitations. First, they only consider the present joint configuration, discarding past information. This approach might lead to inaccuracies in predictions as the weight distribution is influenced not only by the current posture, but also by the velocity and acceleration of the motion. Moreover, few considerations in these studies are given to the generalization across subjects and the prediction time performances. The latter is essential for real-time implementation of the exoskeleton’s whole-body controller. To address the aforementioned limitations, we present a deep-learning approach for estimating the weight distribution of a user wearing a lower-limb exoskeleton. This approach utilizes a short time window of kinematic data extracted from sensors integrated in the exoskeleton. For the validation of predictions, we compared the proposed approach with the method in [14] which only uses the instantaneous posture. Employing distinct users for training and testing resulted in a user-independent estimation, eliminating the need for a calibration phase. The proposed deep-learning method was trained on walking data from six users, and its closed-loop performance was evaluated on three users (two of which were not included in the training set) during treadmill and overground walking. The results showed that weight distribution can be accurately estimated using kinematic data with an R 2 of 0.9 averaged across users. In addition, this approach was compatible with real-time requirements, achieving a prediction time less than 1 ms. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Weight Distribution Estimation in Lower-Limb Exoskeletons via Deep Learning: Introduction

Have you ever found yourself staring at an image spotted somewhere on the internet and wondering whether it results from human or artificial creative labor? Since the breakthrough in image generation technology, I’ve been captivated by interior designs, nature scenery, fashion photographs, and futuristic cityscapes made with AI that look like professional artwork. As someone with a creative eye yet no artistic skills, I was excited about the opportunity to transform ideas from my mind into visuals based solely on text input. I tried out several popular text-to-image tools, like MidJourney and DreamStudio on the web and ARTA on mobile. Still, while they all provided a good starting point and a straightforward user experience, the generation outcomes often fell short of my expectations and differed from what I envisioned — until I learned to tell AI what to draw in the language it understands. If you’re new to AI-generated art and need assistance streamlining the creative process, I’ll provide a step-by-step guide below to help you compose concise and well-structured text prompts that the AI models can effectively interpret and translate into the desired images. In the bottom line, I will share an example of both a correct and an incorrect prompt to illustrate how significantly they can impact the generation output in practice. First things first, two common prompt-writing mistakes you should avoid when using any text-to-image generator are being too wordy and using complex grammar. In fact, the more words you incorporate and the more complicated your grammar, the more difficult it is for AI to understand your prompt and adequately depict your concept. Therefore, communicate concisely, carefully choosing keywords, and list your ideas explicitly and separately rather than trying to cram them into one sentence. Step 1. Choose the Content Type Ask yourself how you want your image to look overall. Do you see it as a photograph, illustration, painting, sketch, or other artwork type? This initial decision puts you in control of the final result and helps align all the elements with this general framework. For instance, choosing a sketch will typically produce pencil images, while opting for a painting will add colors and likely create a more vintage appearance. Step 2. Define the Subject AI image generators require a specific subject for rendering, so you must indicate the main subject of your future image, whether a person, an animal, or a landscape. Stick to one subject per prompt and opt for concrete nouns like girl, dog, train, rose, ocean, temple, etc., for more accurate results. You can most likely render something using abstract keywords like love, peace, or joy, but the results may appear very inconsistent in what they depict. Step 3. Add Subject Description Now, it’s time to detail the concept with modifiers – adjectives or attributive nouns that add to the sense of a head subject. Modifiers can completely change the perspective of your image and boost its overall quality, so the more descriptive you are, the better the generated output. Define what the subject looks like, what the subject is doing, and other characteristics essential to representing your concept. Step 4. Describe the Context To help AI create the exact atmosphere and surroundings you need, give attention to aspects like lightning (studio, soft, cinematic, volumetric, accent lighting, sunlight, frontlight, backlight, ambient, moonlight, etc.), color theme (pastel, vibrant, neon pink, dark, etc.), environment/scene (in the sky, indoor, underwater, at night, etc.), and background (forest, beach, solid color, nebula, etc.). Additionally, consider the mood and vibe you want to convey. Simple feeling modifiers can set the scene’s atmosphere, whether you opt for positive (energetic, cozy, romantic, satisfaction, etc.) or negative (depressive, loneliness, regret, fear, etc.) ones. Step 5. Define Perspective Add a point of view to determine the angle from which you are looking at the image, such as side, overhead, or front. Other essential modifiers for creating photo-like images involve lenses and cameras. Here are a few examples: close-up, macro, lens flare, wide angle, microscopic, bokeh, isometric, depth of field, ultra-wide angle, fisheye lens, and panorama. Step 6. Add Art Style, Art Mediums, and Artists Proceed to aesthetic settings and define the overall artistic style. Also, specify the art medium. These aspects can greatly contribute to the value of the generated output, making it more consistent and expressive. Here are some of the common keywords: geometrical art, pop art, watercolor, line drawing, line art, 3D sculpture, crayon drawing, painting, oil painting, pastel drawing, medieval art, photography, digital art, stained glass, cave painting, polaroid, vector art, retro, illustration, comics, editorial fashion photography, Japanese anime style, surreal, travel photography, pencil sketch, steampunk, cartoon, futurism, vaporwave, art deco, interior design, psychedelic, street art. Consider including artistic influences. Adding the name of a specific artist whose signature artwork aligns with your chosen art genre can enhance the final image. Note, however, that finding the right artist is crucial so as not to mess up the output. Some artists affect the generation only subtly, while others can alter it significantly depending on their distinctive manner. For example, Claude Monet can infuse impressionism into landscapes, while mentioning Salvador Dali adds a surrealistic touch. Step 7. Add “Magic” Words Certain words can make a dramatic difference in the outcomes of the generation. The keywords like highly detailed, ultra-realistic, 4K, 8K, UHD, and HDR are the game changers for the overall quality of the image. Adding “studio lighting” will introduce appealing textures to the image, while “accent lighting” can draw attention to specific features. Describing the image as “professional” can enhance color contrast and detail, while mentioning “vivid colors” will bring life to your image. Using terms like ultra-wide angle, panoramic, or long-shot will help guide the camera. Packing your prompt with phrases related to artistic techniques and materials (such as trending on artstation, precise line-art, golden ratio, fine art, etc.) and those associated with artistic value (such as most beautiful image ever seen, epic composition, etc.) can also bring sophisticated improvements. Step 8. Consider Adding Negative Prompts Negative prompts are parameters that inform the generation model about what you do not want to see in the final image. They help prevent AI from generating specific elements, fix image abnormalities, and improve overall quality. Here are some universal examples: blurry, cropped, cloned face, duplicate, error, ugly, wrong proportions, deform, fused fingers, extra arms, law quality res, watermark, and more. Step 9. Define Correct Word Order and Keyword Weight Input The order is almost as important as the vocabulary, guiding the AI in generating the desired output. Combine keywords to direct the focus of your prompt and note key phrases that influence image preference. Emphasize essential keywords with a full colon and decimal number, use parentheses for added importance, and brackets for less importance. Typically, the earlier a word appears in the sentence, the more importance it is given. If the AI model didn’t take in some key phrase and failed to depict it in the generated image, try rearranging the order of words, placing the missing element earlier in the prompt. Bad vs. Good Prompt Example Bad prompt: portrait of a woman looking directly at the camera, neck tattoo, urban pink background. Good prompt: head and shoulders realistic photo portrait of a beautiful woman looking at the camera, white skin, pink lips, red hair, blue eyes, black tattoos on her neck, neon punk, highly detailed, pink neon city background. Creating effective prompts requires both artistic imagination and technical skills. By including the necessary building blocks and following the proper prompt structure you ensure that the text-to-image generation model understands your vision and produces optimal output. However, don’t be afraid to experiment; the more you practice, the more adept you become, and the more remarkable your creations will be. Have you ever found yourself staring at an image spotted somewhere on the internet and wondering whether it results from human or artificial creative labor? Since the breakthrough in image generation technology, I’ve been captivated by interior designs, nature scenery, fashion photographs, and futuristic cityscapes made with AI that look like professional artwork. Have you ever found yourself staring at an image spotted somewhere on the internet and wondering whether it results from human or artificial creative labor? Since the breakthrough in image generation technology, I’ve been captivated by interior designs, nature scenery, fashion photographs, and futuristic cityscapes made with AI that look like professional artwork. As someone with a creative eye yet no artistic skills, I was excited about the opportunity to transform ideas from my mind into visuals based solely on text input. I tried out several popular text-to-image tools, like MidJourney and DreamStudio on the web and ARTA on mobile. Still, while they all provided a good starting point and a straightforward user experience, the generation outcomes often fell short of my expectations and differed from what I envisioned — until I learned to tell AI what to draw in the language it understands. If you’re new to AI-generated art and need assistance streamlining the creative process, I’ll provide a step-by-step guide below to help you compose concise and well-structured text prompts that the AI models can effectively interpret and translate into the desired images. In the bottom line, I will share an example of both a correct and an incorrect prompt to illustrate how significantly they can impact the generation output in practice. First things first, two common prompt-writing mistakes you should avoid when using any text-to-image generator are being too wordy and using complex grammar. In fact, the more words you incorporate and the more complicated your grammar, the more difficult it is for AI to understand your prompt and adequately depict your concept. Therefore, communicate concisely, carefully choosing keywords, and list your ideas explicitly and separately rather than trying to cram them into one sentence. Step 1. Choose the Content Type Ask yourself how you want your image to look overall. Do you see it as a photograph, illustration, painting, sketch, or other artwork type? This initial decision puts you in control of the final result and helps align all the elements with this general framework. For instance, choosing a sketch will typically produce pencil images, while opting for a painting will add colors and likely create a more vintage appearance. Step 2. Define the Subject AI image generators require a specific subject for rendering, so you must indicate the main subject of your future image, whether a person, an animal, or a landscape. Stick to one subject per prompt and opt for concrete nouns like girl, dog, train, rose, ocean, temple, etc., for more accurate results. You can most likely render something using abstract keywords like love, peace, or joy, but the results may appear very inconsistent in what they depict. Step 3. Add Subject Description Now, it’s time to detail the concept with modifiers – adjectives or attributive nouns that add to the sense of a head subject. Modifiers can completely change the perspective of your image and boost its overall quality, so the more descriptive you are, the better the generated output. Define what the subject looks like, what the subject is doing, and other characteristics essential to representing your concept. Step 4. Describe the Context To help AI create the exact atmosphere and surroundings you need, give attention to aspects like lightning (studio, soft, cinematic, volumetric, accent lighting, sunlight, frontlight, backlight, ambient, moonlight, etc.), color theme (pastel, vibrant, neon pink, dark, etc.), environment/scene (in the sky, indoor, underwater, at night, etc.), and background (forest, beach, solid color, nebula, etc.). Additionally, consider the mood and vibe you want to convey. Simple feeling modifiers can set the scene’s atmosphere, whether you opt for positive (energetic, cozy, romantic, satisfaction, etc.) or negative (depressive, loneliness, regret, fear, etc.) ones. Step 5. Define Perspective Add a point of view to determine the angle from which you are looking at the image, such as side, overhead, or front. Other essential modifiers for creating photo-like images involve lenses and cameras. Here are a few examples: close-up, macro, lens flare, wide angle, microscopic, bokeh, isometric, depth of field, ultra-wide angle, fisheye lens, and panorama. Step 6. Add Art Style, Art Mediums, and Artists Proceed to aesthetic settings and define the overall artistic style. Also, specify the art medium. These aspects can greatly contribute to the value of the generated output, making it more consistent and expressive. Here are some of the common keywords: geometrical art, pop art, watercolor, line drawing, line art, 3D sculpture, crayon drawing, painting, oil painting, pastel drawing, medieval art, photography, digital art, stained glass, cave painting, polaroid, vector art, retro, illustration, comics, editorial fashion photography, Japanese anime style, surreal, travel photography, pencil sketch, steampunk, cartoon, futurism, vaporwave, art deco, interior design, psychedelic, street art. Consider including artistic influences. Adding the name of a specific artist whose signature artwork aligns with your chosen art genre can enhance the final image. Note, however, that finding the right artist is crucial so as not to mess up the output. Some artists affect the generation only subtly, while others can alter it significantly depending on their distinctive manner. For example, Claude Monet can infuse impressionism into landscapes, while mentioning Salvador Dali adds a surrealistic touch. Step 7. Add “Magic” Words Certain words can make a dramatic difference in the outcomes of the generation. The keywords like highly detailed, ultra-realistic, 4K, 8K, UHD, and HDR are the game changers for the overall quality of the image. Adding “studio lighting” will introduce appealing textures to the image, while “accent lighting” can draw attention to specific features. Describing the image as “professional” can enhance color contrast and detail, while mentioning “vivid colors” will bring life to your image. Using terms like ultra-wide angle, panoramic, or long-shot will help guide the camera. Packing your prompt with phrases related to artistic techniques and materials (such as trending on artstation, precise line-art, golden ratio, fine art, etc.) and those associated with artistic value (such as most beautiful image ever seen, epic composition, etc.) can also bring sophisticated improvements. Step 8. Consider Adding Negative Prompts Negative prompts are parameters that inform the generation model about what you do not want to see in the final image. They help prevent AI from generating specific elements, fix image abnormalities, and improve overall quality. Here are some universal examples: blurry, cropped, cloned face, duplicate, error, ugly, wrong proportions, deform, fused fingers, extra arms, law quality res, watermark, and more. Step 9. Define Correct Word Order and Keyword Weight Input The order is almost as important as the vocabulary, guiding the AI in generating the desired output. Combine keywords to direct the focus of your prompt and note key phrases that influence image preference. Emphasize essential keywords with a full colon and decimal number, use parentheses for added importance, and brackets for less importance. Typically, the earlier a word appears in the sentence, the more importance it is given. If the AI model didn’t take in some key phrase and failed to depict it in the generated image, try rearranging the order of words, placing the missing element earlier in the prompt. Bad vs. Good Prompt Example Bad prompt: portrait of a woman looking directly at the camera, neck tattoo, urban pink background. Good prompt: head and shoulders realistic photo portrait of a beautiful woman looking at the camera, white skin, pink lips, red hair, blue eyes, black tattoos on her neck, neon punk, highly detailed, pink neon city background. Creating effective prompts requires both artistic imagination and technical skills. By including the necessary building blocks and following the proper prompt structure you ensure that the text-to-image generation model understands your vision and produces optimal output. However, don’t be afraid to experiment; the more you practice, the more adept you become, and the more remarkable your creations will be.

How to Write Text Prompts for AI Image Generators to Get Optimal Results with Ease

Simplifying Transformer Models for Faster Training and Better Performance

The Future of AI in Finance

Stories, News & Mentions about SOTAI

Sotai’s Stories on HackerNoon