THIRDAI: SQUEEZING MORE OUT OF CHIPS USING ALGORITHMS AND SOFTWARE INNOVATIONS. ThirdAI (pronunciation: /THərd ī/ Third eye) is a cutting-edge Artificial intelligence startup carving scalable and sustainable AI. UNIQUE APPROACH- ThirdAI accelerator builds hash-based processing algorithms for training and inference with neural networks. The technology is a result of 10 years of innovation in finding efficient (beyond tensor) mathematics for deep learning. Our algorithmic innovation has demonstrated how we can make Commodity x86 CPUs 15x or faster than most potent NVIDIA GPUs for training large neural networks. The demonstration has shaken the common knowledge prevailing in the AI community that specialized processors like GPUs are significantly superior to CPUs for training neural networks. Several media outlets echo our technological superiority: IEEE Spectrum describes, "What they report is stunning." Investorplace considers the algorithm one of the biggest threats to NVIDIA Stock (a $500 billion company). Engadget calls it a "Completely different approach to Deep Learning". A quote from WCCFTech says, "If an implementation of this algorithm is mainstreamed it would almost instantly disrupt the dynamics of the deep learning ecosystem. Valuations of companies could change overnight." Our innovation would not only benefit current AI training by shifting to lower-cost CPUs, but it should also allow the “unlocking” of AI training workloads on GPUs that were not previously feasible.

ThirdAI

People have countless fantasies about Artificial Intelligence. It has become the most popular theme in novels and movies. When we dream about AI, we often fancy a world with Iron Man and his intelligent assistant J.A.R.V.I.S (or it’s replacement FRIDAY); Baymax from Big Hero 6; or the high-tech adult theme park from Westworld. However, due to the current technological advancement, the humanoid robots that transcend our human ability is still far away from us. Artificial Intelligence tools nowadays serve as assistants to our lives and solve pain points that could not be addressed before. So, what is preventing the birth of humanoid robots? First, let’s take a brief look at the history of AI. The field of AI was formally established at a conference held at Dartmouth University in 1956 and coined the term “Artificial Intelligence.” Between 1956 and 1974, universities and government agencies invested a countless amount of money and resources in AI researches. Different types of computers were invented that were solving algebra, geometry, and so on. The success of this period made people overly optimistic and began to have high expectations for AI. However, the technology and hardware equipment at the time could not keep up with the novel AI, leading to the first AI winter started in 1974. The winter continued until the early 1980s, due to the rise of the expert systems, coupled with Japan’s $850 million attempt to create a computer with supercomputing power with human intelligence, known as the Fifth generation computer project. Japan has long been obsessed with humanoid robots. The widely-known science-fiction franchise, Gundam (humanoid vehicles with advanced AI systems), was started in 1979. The original Ghost In The Shell comic was also first released in 1989. To compete with Japan, the British government also invested in AI again during the same period. During this period, computer hardware boomed. The computers mentioned here were not the kinds known to the average joes but used by professionals with no interface and only a bunch of codes. In 1987, Apple and IBM both developed their own versions of desktop computer (similar to the ones we are using today) with higher performance than those professional computers. This has led to the collapse of the $500 million expert system industry leading to the second AI winter. It wasn’t until 1997 when IBM’s Deep Blue defeated chess champion, Garry Kasparov, that made the general public to realize the power of AI. At the same time, the tech industry faced the Dot-com bubble. Funds invested in AI were exhausted. However, machine learning continued to advance, mainly due to the breakthrough in hardware equipment. The computer’s processing power and storage capacity grew exponentially, enabling companies to store large amounts of data and process large amounts of data. Various companies and government agencies have successfully applied AI on a larger scale in different applications. In 2011, IBM’s question-and-answer system Watson defeated reigning champions Brad Rutter and Ken Jennings in the American quiz show “Jeopardy!”. In 2016 and 2017, Alpha GO defeated Go player of 9 dan rank, Lee Sedol, and the world champion, Ke Jie, respectively. For the past 15 years, Amazon, Google, and other companies have gained substantial competitive advantages using machine learning. In addition to processing user data to understand consumer behavior, these companies continue to focus on computer vision, natural language processing, and many other AI applications. After decades of research, four key factors have contributed to the rapid development of AI: Moore’s Law — Since the beginning of AI research, computer processing power has approximately doubled every two years to the point that handling big data and complex algorithms are made possible. It only took Google two years to make Alpha Go from an amateur Go player to defeat the world champion. Also, the latest version of Alpha Go’s computing resources is only one-tenth of the previous version that defeated Lee Sedol; Big Data — Machine learning relies on a large amount of raw data. Over the last two years alone has generated 90 percent of the data in the world; Funding — As AI becomes more widely adopted, VC and angel funds are pouring into the AI research, hoping to find the next unicorn; Algorithm Complexity — Research and funding are driving increasingly sophisticated algorithms that is possible to utilize neural networks and cognitive computing fully. Many people believe that the current development of AI has slowed down to the point that we are near to face the third AI winter. Some think that there has been no breakthrough in the development of strong AI. Others believe that today’s enterprises focus on weak AI applications which suppress the growth of the entire AI industry. There are also people who criticize that the current that is harming society. AI research is too centralized However, the current stock market is dominated mainly by the likes of Apple, Google, Amazon, Facebook, and so on. If the third AI winter were to come, not only will the AI industry stop moving forward, but the entire world will also face the next economic crisis.

Is The Third AI Winter Coming?

With all the assumptions we make about the advancements in enterprise data and cloud technologies, there’s a plain fact that often gets overlooked: The majority of the most important enterprise data remains in the corporate data center. There are plenty of reasons for this — some reasonable, some not so much. In some cases, it’s because of the highly sensitive nature of data, whether it’s HIPAA compliance, sensitive banking data or other privacy concerns. In other cases, the data resides in systems (think legacy enterprise resource planning data or petabyte-scale scientific research data) that are difficult to move to the cloud. And sometimes, it’s just inertia. It’s not a great excuse, but it happens all the time. Whatever the reason, housing data on racks of corporate servers has proved to be a real hindrance to many enterprises’ ability to take advantage of AI to transform their business, because it’s been all but impossible to provide the significant compute power necessary to drive AI on the infrastructure underpinning most data centers. But there’s a movement under way, via a small constellation of startups and big device makers, to optimize machine learning models and make AI available to companies whose data isn’t in the cloud. It’s going to be a game changer. The processing power problem The graphical processing unit, or GPU, was developed to handle high-intensity video-processing applications like those required by modern video games and high-resolution movies. But the ability of these processors to break down complex tasks into smaller tasks and execute them in parallel also makes these high-powered application-specific integrated circuits (ASICs) very useful for artificial intelligence. AI, after all, requires massive streams of data to refine and train machine learning models. CPUs, on the other hand, are the flexible brains of servers, and, as such, they are built to handle a wide variety of operations, like accessing hard drive data, or moving data from cache to storage, but they lack the ability to do these tasks in parallel (multicore processors can handle parallel tasks, but not to at the level of GPUs).They simply aren’t built to handle the kind of high-throughput workloads that AI demands. High-performance GPUs are very expensive and until recently, they’ve been scarce, thanks to the reliance of crypto miners on these high-performance chips. For the most part, they’re the realm of the cloud providers. Indeed, high-performance computing services are a big reason companies move their data to the cloud. Google’s , or TPU, is a custom ASIC developed solely to accelerate machine learning workloads. Amazon also has its own chips for powering AI/ML workloads. Tensor Processing Unit Optimizing for AI GPUs have been the foundation of the rush of AI innovation that has recently taken over the headlines. Much of these high-profile developments have been driven by companies pushing the envelope on what’s possible without thinking too much about efficiency or optimization. Consequently, the workloads produced by new AI tools have been massive, and so, by necessity, managed in the cloud. But in the past six months or so, that’s been changing. For one thing, the sprawling ML models that drive all of these cutting-edge AI tools are getting condensed significantly, but still generating the same powerful results. I installed the app on my mobile phone, for example. It’s a 13 billion-parameter model that does ChatGPT-like execution and runs in real time, right on my phone. It’s not in the cloud at all – it’s an app that resides on a device. Vicuna The Vicuna project emerged from the , a collaboration between UC Berkeley, UC Davis and Carnegie Mellon University with the mission to “make large models accessible to everyone by co-development of open datasets, models, systems, and evaluation tools.” Large Model Systems Organization It’s a mission that big tech isn’t ignoring. Apple’s latest desktops and iPhones have specialized processing capabilities that accelerate ML processes. Google and Apple are doing a lot of work to optimize their software for ML too. There’s also a ton of talented engineers at startups that are working to make hardware more performant in a way that makes AI/ML more accessible. is a great example. The company offers a software-based engine that can train large deep-learning models by using CPUs. DataStax has been experimenting with the ThirdAI team for months and has been impressed with what they have developed — so much so that last week we announced a partnership with the company to make sophisticated large language models (LLMs) and other AI technologies accessible to any organization, regardless of where their data resides. (Read more about the partnership news ). ThirdAI here Bring AI to the data Because of all this hard work and innovation, AI will no longer exclusively be available to organizations with data in the cloud. This is extremely important to privacy, which is a big reason many organizations keep their data on their own servers in the first place. With the AI transformation wave that has washed over everything in the past 18 months or so, it’s all about data. Indeed, there is no AI without data, wherever it might reside. Efforts by teams like ThirdAI also enable all organizations to “bring AI to the data.” For a long time, companies have been forced to do the opposite: bring their data to AI. They had  to dedicate massive resources, time and budget to migrate data from data warehouses and data lakes to dedicated machine learning platforms before analyzing for key insights. This results in significant data transfer costs, and the required time to migrate, analyze, and migrate affects how quickly organizations can learn new patterns and take action with customers in the moment. Bringing AI to the data is something we have focused on a lot at DataStax with our , because it’s the fastest way to take actions based on ML/AI, delight customers and drive revenue. Bringing AI to the data center, and not just the cloud, is another very important step to making the transformational AI technology wave something all companies can be a part of. real-time AI efforts *Learn about the new which connects enterprises with groundbreaking AI startups to accelerate the development and deployment of AI applications for customers.* DataStax AI Partner Program , By Ed Anuff, DataStax Also published . here

Bringing AI to the Datacenter

Llevando la IA al centro de datos

Apporter l'IA au Datacenter

डाटासेंटर में एआई लाना

AIをデータセンターに導入

Trazendo IA para o Datacenter

Đưa AI vào Datacenter

将 AI 引入数据中心

AI is flourishing with the rise of ChatGPT, while crypto crashes abound. So why can’t I stop thinking about blockchain? Artificial Intelligence (AI) is currently tech’s golden child. AI-powered tools like ChatGPT are now sophisticated enough to pass , and almost reported using ChatGPT to draft emails or code. MBA and medical license exams 30% of surveyed professionals Meanwhile, sentiment around blockchain is pessimistic. Justifiably so, after countless people lost their life savings amidst price crashes and the implosion of centralized cryptocurrency exchanges. I’ve worked in both spaces. I started as a product manager within Google’s Research & Machine Intelligence division, developing a product that I transitioned completely by accident when I discovered diagnosed patients with preventable eye diseases. blockchain could be leveraged for fundraising. Given the current momentum behind AI applications and the steep headwinds against blockchain, I’ve been asked repeatedly why I ever made the switch. The answer is simple: While AI empowers individuals, blockchain can empower collectives. Tech isn’t interesting without a concrete application: a hammer seems pretty lame until it’s used to build a house. So what are the applications of AI and blockchain respectively? In short: Let’s break this down with real-world examples that I’ve experienced. Artificial Intelligence: Empowering Individuals At Google, I worked on a product that assessed a patient’s risk of blindness based on the image of the back of their eye. Doctors used the tool below to generate a data set of 100k+ images and their associated disease status: After training on this data set, the AI could diagnose patients with . While the AI performed at par with expert physicians, it boasted two significant improvements: ~90% accuracy The model outputs a diagnosis in seconds, while humans take ~10 minutes, increasing the volume of diagnosed cases by 100x. Speed: The AI consistently outperformed less experienced and overextended physicians, since it trained on a significantly larger data set and is immune to fatigue Quality: This use case highlights the key value proposition of artificial intelligence: to replace or expedite human labor that is routine and/or error prone. The impact of introducing AI to this workflow is clear: by handing over the time-consuming, repetitive task of image recognition to AI, physicians could focus on what they uniquely do best: patient interaction. Rather than humans, we found that AI individuals through the time and effort saved of handling lower-level tasks. replace empowers ChatGPT promises to empower a broader set of individuals. Writers and developers who leverage ChatGPT will let the computer handle the boring stuff, creating more room for creative thinking. I’ve personally experienced this shift, saving energy by generating this post’s splash image on and leveraging ChatGPT to SEO-optimize my prose. MidJourney I’m a big believer in the power of AI to empower individuals and planned to continue my tech career in the space. So what pulled me away? My Unexpected Encounter with Blockchain: Museum Volunteering While I was clear on the value prop of AI when I joined Google, my foray into blockchain could not have been more random. I never understood the point of cryptocurrency, accusing my dad of gambling when he gifted me ETH in 2018 and selling it all when I thought it peaked at $28 (oops). But during COVID, I began volunteering for a museum. , a trend I hoped to reverse as an art lover. As I explored strategies for non-profit fundraising, I came across an unexpected solution: blockchain. A third of U.S. museums were at risk of closing permanently If I donate to a museum (or any cause) today, I have no idea whether the money is allocated as promised. I also have no say over how the funds are distributed since that is left entirely to the organization. Blockchain disrupts this model. As a quick refresher, forget any of the overcomplicated jargon you’ve heard to define blockchain. Blockchain is just like any other database (record of entries, such as transactions), except that no single entity has control over it (de-centralized). Traditional fundraisers track transactions on centralized databases, controlled by banks or the fundraiser. These databases are managed by and only accessible to their owners, meaning donors have no visibility. In contrast, records on public blockchains are viewable by anyone, holding fundraisers accountable for the proper allocation of the sum they raise. Due to its decentralized nature, blockchain ensures that fundraisers are transparent. Pushing fundraising entities towards transparency already represents a significant improvement over the status quo. But the true power of blockchain emerges when there is no central fundraising entity at all. A Donor-Managed Fundraiser Imagine your hometown is hit by a natural disaster, leaving thousands without access to food or water. You want to help, but you want your dollars to be spent on initiatives you believe in. In the status quo, your only option is to donate to a fundraising entity (e.g. non-profit, campaign) and cross your fingers that they account for your opinion, and they use the money as promised. That requires putting a lot of faith in a group of people you’ve never met. a) b) But what if you could donate to a fundraiser that was managed entirely by its donors, rather than a third party? Logic is built into the blockchain that distributes the funds based on the outcome of votes. Beyond providing a transparent and secure database, many blockchains support an additional feature: voting rights. automatically For example, a blockchain-powered fundraiser may use a majority vote to evaluate proposals for how to allocate collected funds. A donor proposes paying Company X to rebuild homes. Another donor proposes hiring Company Y. If the majority votes for Company Y, then the funds tracked on the blockchain are automatically disbursed  to Company Y. Funds thereby pass from donor to operator without passing through any intermediary, while automatically accounting for the input of donors. As a donor, you no longer need to rely on a fundraising entity to ensure that your donations are properly allocated. This use case highlights the utility of blockchain to coordinate people without needing to trust one another. Blockchain: Empowering Collectives I’m still extremely skeptical of most blockchain use cases and products. Unregulated speculation on cryptocurrency and NFTs have destroyed countless lives, and most blockchain products feel like they are built cyborgs other cyborgs. The space requires a massive rehaul of values and user experience before it has any chance of achieving impact and adoption at scale. by for Despite my doubts, my experience with blockchain-powered fundraising gave me valuable insight in the technology’s latent potential: by automatically and transparently handling tasks that require a high level of trust (e.g. fund disbursement), blockchain empowers collectives to manage themselves without relying on a third-party. Since the database tracking activity (e.g. transactions, voting) of the group is decentralized, and because voting outcomes are automatically enforced, collectives can self-organize without needing to find a trustworthy “leader” or administrator. Pretty f**king cool! Now that I’ve worked on both (Blockchain and AI) — what’s next? There is no way to predict whether Blockchain vs. AI will have a greater societal impact. They are each mere tools, and their legacy is entirely dependent on how the tools are leveraged (or abused). But blockchain doesn’t stand a chance if the tech community continues to fixate on its negative applications while obsessing over AI. The meaningful applications of blockchain will never see the light of day if builders are scared away by this distorted sentiment. Given my conviction that blockchain is a powerful tool to organize collectives, I plan to continue sharing perspectives on the potential of the technology when used correctly. So along with the community of writers at , I’ll keep covering meaningful applications of blockchain — stay tuned on and for my next post! Bankless Academy Twitter Medium Notes: For the sake of non-jargony writing, I’m abstracting away the concept of smart contracts and DAOs here. This assumes that Company Y has a cryptocurrency wallet! Or in the future, smart contracts can interact with traditional bank accounts. Technical note: Featured image generated by MidJourney.

Why I Left Google's AI Division for the World of Blockchain

কেন আমি ব্লকচেইনের জগতের জন্য গুগলের এআই বিভাগ ছেড়েছি

Warum ich die KI-Abteilung von Google verlassen habe, um in die Welt der Blockchain zu wechseln

Por qué dejé la división de IA de Google por el mundo de Blockchain

Pourquoi j'ai quitté la division IA de Google pour le monde de la blockchain

मैंने ब्लॉकचेन की दुनिया के लिए Google के AI डिवीजन को क्यों छोड़ा

ブロックチェーンの世界に向けて Google の AI 部門を離れた理由

내가 블록체인 세계를 위해 Google AI 부서를 떠난 이유

Por que deixei a divisão de IA do Google pelo mundo do Blockchain

Blockchain Dünyası İçin Google'ın Yapay Zeka Bölümünden Neden Ayrıldım?

Tại sao tôi rời bộ phận AI của Google để đến với thế giới chuỗi khối

为什么我离开谷歌的人工智能部门进入区块链世界

Artificial intelligence is starting to hit its stride, and many obstacles along the way have been addressed—the availability of computing power, data management systems, and so on. However, there are still many challenges ahead for the to be successful. top industry trends for AI marketing Here are some of the most significant challenges you’re likely to encounter and how you can mitigate them. Data Acquisition A fundamental challenge for startups using AI is acquiring enough first-party customer data from their users. It is hard to be explicit about how much data a company needs to truly leverage AI—that depends on what use cases they want to start with. This could be based on conversion goals or historical conversion rates; for example, this could be for analytics or segmentation and targeting. I like to think about artificial intelligence and machine learning as a tool chest to help get the benefits of . With things like deep learning, the tool chest just got deeper and has more powerful tools. marketing automation Therefore, all companies, from early-stage startups to big multibillion-dollar businesses, can leverage AI provided they have the correct data acquisition strategy. I think it’s fair to say that startups have to be even more strategic in what data they collect and how they collect it; for example, in some cases, they can bootstrap their algorithms by licensing third-party data sets to start with—and improve the performance of their models with their first-party data. Most of the friction in data collection results from customers’ lack of trust in startups that lack brand awareness or do not provide a compelling value proposition for solving their problems. It is critical to building trust and a strong relationship quickly when you’re asking for data from new customers, and the way to do that is to reiterate value consistently. Data capture comes even before the algorithms to establish long-term success in leveraging AI and machine learning in growth marketing. The value of first-party customer data is priceless, and no publicly accessible data will ever provide the same competitive edge. Access to unique data isn’t a problem for major platforms like Google, Facebook, and Amazon because they are well-established brands that continue to provide clear value in return for capturing user data. However, most startups are beginning at ground zero to convince people to share their data with them. Unfortunately, AI is not useful until enough people are using your product or service for you to capture a critical mass of data events. Only with this flow of information can the AI intelligent machine help with your acquisition, retention, and monetization efforts across the entire customer journey. Getting the data flywheel process started and keeping it going is something all businesses face with AI, but this is true even more for startups. However, startups should not give up on AI because, in the long run, it offers many opportunities to disrupt, innovate, and solve problems faster. The key is to ensure you have a good product/market t that compels people to give up their data before you have reached the scale to provide optimal value from the data flywheel kicking in. The value of first-party customer data is priceless, and no publicly accessible data will ever provide the same competitive edge. This means you need to invest the effort to develop a robust data acquisition strategy that provides value or incentives to customers for giving up their data. An example is Amazon Prime, which offers much more than free two-day shipping to members, all because Amazon knows that Prime members have significantly higher LTV than non-Prime members. Because of the value add, Amazon captures a wealth of customer data that enables it to better target new users and make product recommendations to current customers. The key to their success was investing in a data acquisition strategy from the beginning, then using that data to give value back to their customers consistently. Privacy Controls Data is a commodity that will continue to increase in value as more data regulations like GDPR and privacy controls are adopted worldwide. I would expect more major media platforms like Google, Facebook, and others to use proactive “privacy” measures to give users more transparency about what data they want to share with advertisers. It’s hard to predict how this will impact these companies regarding their customer data policies, but it will surely be a major distraction for them. It could lead to more challenges for growth if the companies end up sharing less data to identify and track users coming from their platforms, where the bulk of the paid acquisition budget is spent. There is a big risk that Apple and Google may completely stop sharing mobile device app IDs with the attribution platforms, which play a collection in tracking customers on mobile devices. This could lead to AI intelligent machines being forced to make decisions with less accuracy and transparency in the attribution data they depend on. Cross-platform attribution about the customer journey is already challenging, but it will get even more so shortly as these major technology companies are scrutinized on all fronts regarding user privacy. Facebook's privacy policies and all the other major media partners will be an ongoing challenge as they evolve. They all want to empower users to influence the regulatory debate around data privacy. The truth is, by requiring users to make many changes to opt out, most people will ignore it and end up using the Facebook and Google default permission settings (which will let the companies scoop up their information). Team Downsizing There are plenty of scary headlines in the news about AI killing off jobs. A new report by that by 2030 as many as 800 million jobs could be most customers highly value personalization recommendations, content, and offers most customers highly valued personalization recommendations, content, and offers that affect everyday working lives, comparable to the shift away from agricultural societies during the Industrial Revolution. In the United States alone, between 39 and 73 million jobs are expected to be automated—making up around a third of the total workforce. McKinsey Global Institute predicts Businesses will see transformational change in all areas of their organizational structure. This will require them to retool their business processes and reevaluate their talent strategies and workforce needs. They will have to carefully consider which workers are needed, which can be redeployed to other jobs, and where new talent is required. There is a danger of a political backlash if unemployment goes up. Many companies are finding it is in their self-interest and part of their societal responsibility to train and prepare workers for a new world of work. There will be an impact on the growth team as more roles and tasks get automated with AI. This will be especially evident on bigger growth teams where many campaigns, media buying, and data scientist roles become obsolete as machines prove they can handle that work better, smarter, and more efficiently than humans. If you know AI will be a part of your future growth strategy; you have to make sure your team is aware of this and willing to learn how it will improve their jobs. It’s essential to invest time and resources into training employees who will be key to the success of the AI intelligent machine. Future growth teams will be much leaner as the organizational structure evolves for humans and machines to coexist. Machines will become much smarter, more productive, and achieve better results leveraging artificial intelligence, with humans playing the supporting role to empower them. This process will start with the automation of small tasks but will scale up over time. Future growth teams will be much leaner as the organizational structure evolves for humans and machines to coexist. The best safeguard for workers is to take proactive action to build up their skillset so they are relevant in the future of work. The growth marketing team has many roles with rudimentary tasks that are ripe for automation. Instead of worrying about job losses, spend time acquiring new and relevant skills that will allow you to perform higher-level tasks in the technical, strategic, creative, problem-solving, communication, and leadership arenas. AI will create a demand for new jobs that will benefit workers if they stay open to developing the needed talents and capabilities. New Channels and Opportunities There is an old saying that is very appropriate for this dawning age of AI— “don’t put all your eggs in one basket.” In other words, a business should use multiple channels and diversify across different paid and organic platforms to maximize its ability to acquire customers. The key is to reduce risk by not being highly dependent on any one source of traffic. A mistake many startups make is to use paid user acquisition teams that focus on a few highly dependable channels like Google and Facebook. Both are good-quality traffic sources, but your startup paid customer acquisition risk is tied to them not imploding, or else you’re done as well. The challenge is to ensure that the different channels you test are easily integrated into your AI intelligent machine. The best way to do that is to make sure that they have APIs and are already set up within your attribution measurement partner, so you can easily manage new campaigns by adjusting the business inputs (bids, budgets, creative, and goals) to test, learn, and iterate at scale using AI to achieve your desired business outcomes. Staying on Top of Fraud In our modern world, nothing is certain except death, taxes, and fraud. As soon as advertisers caught up with incentivized traffic and bot farm schemes years ago, fraudsters quickly devised new mechanisms to swindle advertisers out of their budgets. According to the eMarketer Digital Ad Fraud 2019 report, estimates of fraud vary widely, but even the most conservative estimates put the money involved well into the billions annually worldwide. Recent estimates vary from $6.5 billion to as high as $19 billion, a range that points to the difficulty of measuring fraud’s true impact. Plenty of co-conspirators in the ecosystem are incentivized to keep fraud alive because they are personally benefiting from it. Unfortunately, there is no consistent definition of fraud and no alignment among the key stakeholders in the attribution platforms, ad networks, ad agencies, and media buyers. Ultimately no one is motivated to solve this problem because it would impact their future compensation and revenue. One approach many high-volume advertisers use is to work with a third-party fraud detection tool from their attribution provider to monitor and filter traffic for anomalies. This can be effective because of the sophistication of detection algorithms and their multi-advertiser view of fraudulent traffic. However, the simplest way to minimize fraud is to avoid ambiguous, non-transparent channels and to buy media directly from reputable sources. The ongoing challenge is that fraud impacts attribution by injecting bad data signals into the AI intelligent machine, which plays havoc with the algorithms. Despite technological advances, ad fraud will continue to be a big problem due to the growth of digital ad spending—or, at least, until there isn’t enough easy money to make by the fraudsters. Facing Challenges These challenges are ongoing and ever-changing because AI is still in the infancy stage with a lot of untapped growth in marketing. However, going all in to tackle these ongoing challenges presents many opportunities for learning and growing individually and collectively as a team. At the end of the day, it’s always better to learn to be the disruptor than to be disrupted. About the Author is the Chief Growth Officer of Tynker, with 20 years of experience helping startups grow into successful businesses. Lomit has previously played a critical role in scaling growth at startups, including Roku (IPO), TrustedID (acquired by Equifax), Texture (acquired. by Apple), and IMVU (#2 top-grossing gaming app). Lomit is a public speaker, author, and advisor, with numerous accolades and awards throughout his career, including being recognized as a Mobile Hero by Liftoff. Lomit's book is part of Eric Ries' best-selling "The Lean Startup" series. Lomit Patel Lean AI

Ongoing Challenges With Artificial Intelligence in Growth Marketing

It’s hard to surprise anyone with artificial intelligence and machine learning nowadays. Even though it’s a young technology, the models and algorithms are already capable of completing many tasks. They can do anything from highly personalized customer service to sophisticated and . AI&ML-powered solutions can bring some benefits to virtually any industry. Real estate is definitely one of them. creative paintings There is plenty of duties in real estate where machine learning algorithms can come in handy. For example, you can use it to estimate how the price of a particular property will change in the future. It will help brokers understand the market and plan their strategy based on real data. And we were lucky to work on such a solution. Here is our experience of building a machine learning model for real estate price prediction. Real Estate Industry: Numbers To Pay Attention To Let us share some interesting statistics that show what real estate looks like right now so we can see the big picture: house markets in the United States. (Zillow) New York, San Francisco, and Chicago are one of the least competitive The prices for recently listed homes of the last year and 26.5% — to the same month of 2020. (Realtor) have risen 13.5% compared to March The median home price has reached $357,300 for 2022. (National Association of Realtors) The number of days that houses stay on the market is decreasing: for March 2022, it was 38 days, compared to . (Realtor) 49 days the previous year Finally, the most interesting one: , Gen X is, with 24% of all people who buy new homes. (National Association of Realtors) Millennials are not the biggest homebuyers on the market All these stats and trends play a part when you estimate and predict the price of a house. Now, we are going to move on to the machine learning case we had recently where all the stats came in handy. Real Estate Price Prediction Model: Why Did We Create It? It's hard to estimate a price of a particular house and predict its behavior in the future. Even if this prediction covers only a short period of time. Too many factors can influence the final numbers. The two most influential of them would be: the prices of the recently sold houses with similar parameters dominating market trends We can see from the previous part that both of these can change drastically with time and some measurements can be seasonal. Besides, there are many characteristics of the house itself that also count: the number of bedrooms, the building’s age, the overall condition, the quality of the neighborhood, proximity to shops, schools, and entertainment, and many more. Also, it’s worth mentioning that professional opinion on a price can have several forms: Comparative Market Analysis Real estate professionals look for comparable homes in the area and define the value of a property based on how those houses behaved on the market. Comparable homes are chosen by size, number of rooms, style, and recent sales price. Broker Price Opinion A Broker Price Opinion (BPO) is another option for a person to get a professional opinion on the house. It is usually made by a professional broker who is aware of the local market. This way is common for short sales, foreclosure, or providing buyers and sellers with the listing price. Such an amount of data that a broker should bear in mind while predicting the house price can get extremely voluminous. This task can be tedious even for the most experienced specialists. Besides, there’s still a possibility of a simple human error. That was the exact reason why we started working on this project a client has come to us with: To automate house price prediction and minimize the influence of possible human error. Our main goal was to make a highly accurate machine learning model that will predict the house price in a month with an accuracy of around 85 to 90 percent. How Did We Create It? Now, the main part: What for price prediction, what tech stack we used, and what main challenge we faced during the process. Let’s start with the overall strategy. Our journey proceeded the following way: steps we took to create a machine learning model . Our first data source was the client itself. They provided us with several datasets, however, it wasn’t enough for model training. To solve this issue, we started the research on other sources that could provide us with real estate data. We used several available sources of information related to the real estate market in the U.S., as well as data related to the country’s economic conditions, in order to achieve a more representative data set. Gathering data . To predict the price, we have chosen the following features: Feature engineering historical change in real estate price property location type of house neighbors presence or absence of a pool other nontraditional variables . Hyperparameters are meant to assist the model during the learning process. They are external to the model meaning they are imposed on it and it has no power to change them anyhow. They are also used only during the training itself and do not complement the model itself. This step was taken so we will have something to validate the model's results, control its behavior, and maximize the performance. Hyperparameter tuning . During the whole process of model training, we were constantly assessing and reevaluating the influence and relevance of each variable. These processes were meant to increase the accuracy of the final solution. Studying variables . When we finished the first version of the model, we started the process anew to polish the model and make sure its results are as accurate as possible. Iterating Technology Stack Our main tool was XGBoost, an open-source gradient-boosted decision tree library for machine learning. We used it for the Regression model. Other items on the list included: Python 3.7—programming language Pandas—data cleaning and analyzing Scikit-learn—classification and predictive analytics This toolset has helped us reach the desired accuracy. The Main Challenge Nothing is perfect in this world, and neither was our machine learning development process. For the most part, it was quite predictable and smooth, but at one moment, we faced the challenge of underfitting. Since the initial dataset was not of the best quality and was quite small, it was hard for the algorithm to find hidden trends and count the accurate results. As we already mentioned, the solution was found in third-party data sources. The information we were able to find in public sources has helped us get back on track and train the model correctly. The Results The results were even better than we expected. The client expected accuracy of around 85-90% compared to the real prices. We were able to achieve 91%. Yes, not that much of a difference, but taking into account the circumstances, we couldn’t expect a better outcome. Was It Successful? In a nutshell, yes, it was. The model was definitely not perfect, the initial data was not the best, and the underfitting issue played its part, but it’s a good start. We were able to see that machine learning is a viable technology in real estate and that we can easily continue working on more cases. Besides, price prediction is not the only area to work with: It also can include home hunting and . property evaluation

How We Used Machine Learning to Predict Real Estate Prices

AI ethics has long been a hot-button issue. For some, it’s reduced to a debate about whether AI should be making decisions traditionally reserved for humans. There are misconceptions that AI systems will quickly evolve into superhuman intelligence, or that they’ll be a silver bullet to solve problems for which we as a society don’t yet have answers. For others, it’s figuring out how to operationalize ethics beyond simply articulating a set of ethics principles. – There’s also too much talk about ethics after negative AI incidents are exposed. As a community, we need to do better. As AI proliferates, it’s time to figure out how to operationalize AI ethics. It’s time to turn ideals into actions, and prime your culture and development process for success. Too much is at stake for anything less. As other countries and big tech companies rush to win the AI race, there are global conversations underway to advocate for AI solutions that safeguard human rights, minimize the impact of unintended bias, and advance the social good. All recognize (though not all will heed) the vital need to establish clear standards for trustworthy AI, and uphold them. Embedding AI Ethics into your Culture The first step is to define your . Numerous policy and research organizations have already done the hard work of analyzing and documenting the different standards and ethics documents around the world, which can serve as a good starting point. Once you’ve established your AI ethics principles, what’s next? ethical principles [i] It’s time to align your culture and institutional processes with your ethical principles. Operationalizing ethical AI requires a shift at all levels within an organization to embed and execute against a fundamentally new way of thinking. When your ethical principles are championed by leadership, embodied throughout staff, and in lockstep with your overarching business goals, it’s much easier for your teams to follow suit and make sure the technology measures up. Leaders should take the conversation forward. Have hard discussions about what applications of your technology meet your principles, and what violates them. These types of difficult conversations ensure everyone is on the same page to take a stance against violations of the principles. This is your chance to create channels to escalate risks and build paths for recourse when mishaps do happen. Most importantly, it’s about incentivizing the people on the ground developing the AI to realize the value of this way of building AI, and empowering them to speak up when something isn’t working. When leaders take responsibility for how the ethics and AI message is spread and embraced throughout a culture, they help ensure the systems developed are accountable. Overseeing the build of AI systems to align with your ethical principles is no small feat. Beyond open discourse among stakeholders, organizations need to institute training programs. Managers and developers must know how to identify and flag risks in systems so that they can be appropriately documented and monitored. Even the most careful organization will have instances where an AI-enabled system runs amiss or doesn’t perform as expected. To address this challenge, last year Accenture announced an AI governance guidebook with recommendations on how to embed governance throughout development teams, including the selection of “fire wardens” with the responsibility to escalate issues when they arise. [ii] By creating mechanisms to document and share lessons learned internally about how and why these incidents happen, you will spur the evolution of an ethics-driven culture, and make these mishaps far fewer in between. Data scientists training models and machine learning engineers building AI-powered applications shoulder the burden of putting these principles into practice. So how do you help them? Build in regular checkpoints and processes for adhering to ethical principles throughout the development process. If the first time your data scientists and developers are thinking about whether a system is transparent or accountable is after it’s deployed, it’s too late.  Work in tandem to establish checkpoints throughout model design and application development. By looking at discrete opportunities to both educate stakeholders and build in frequent checkpoints throughout development, you’ll create a culture and mechanisms to embed your ethical principles and meet the standards you’ve established. AI System Design Much of the work to operationalize AI ethics should occur during AI model design and system development. This requires a certain level of effort outside the normal wheelhouse of a data scientist or developer—that’s why it is so important to incentivize and empower these teams through changes to your organization’s culture. There are also steps your teams can take throughout development to ensure your systems are transparent, accountable, governable, and robust. For many organizations, ethics principles point to the ultimate goal of building trustworthy AI, which requires some degree of transparency into how AI reaches decisions or recommendations. In many cases today, the “black box” excuse no longer cuts it. Stakeholders need to be able to check how or why a model reached a decision. Explainability is one tactic to get at the issue of transparency. You can “explain” how a model reaches decisions or translate the process that AI uses to transform data into insight. Some naysayers immediately ruffle at the mention of this approach because it’s not a panacea—not all AI lends itself to explainability, nor should it be explainable. There are tradeoffs to be made between explainability and system accuracy. DARPA, the Department of Defense’s research agency that brought us the internet, has maintained an active research program focused on AI for the past few years. Part of building transparency into AI is consciously making a choice about what level of explainability is acceptable for a specific application, and documenting and owning that. As mentioned, a key step is developing checklists to aid data scientists and developers in thinking about “ethics” questions in the course of their development activities. While many organizations already have processes and checklists in place for their normal software development activities, AI model design and training, or evaluating models’ adherence to ethical principles, remain more ad hoc activities. Research earlier this year published by Microsoft’s Aether Working Group on Bias and Fairness showed that co-developing checklists with data scientists and machine learning engineers to evaluate an AI model or AI-enabled system’s adherence to ethical principles is effective. Something as simple and clear as the checklist provided in their research paper can help formalize and institutionalize the operationalization of ethics principles. [iii] If you build transparency into models from the onset and establish formal checkpoints, you’ll be better able to evaluate ethics concerns throughout the development process. Once a system is deployed, auditability is crucial to operationalizing ethics. At the Fairness, Accountability and Transparency conference this year, researchers from Google and the Partnership on AI published a paper detailing a framework for auditing AI systems prior to deployment. As organizations deploy AI at an enterprise scale, the potential for risk grows—you need to be able to look under the hood and identify who or what caused an incident. Auditable AI during the design process can be complementary to explainable AI, or in exchange for it when it’s not possible. [iv] It’s up to you to determine how you can achieve this with your software development teams—through building an audit function that logs activity assigned to individual users’ API keys, audit teaming exercises, third-party audits, or bug bounties. There are many options to leverage to proactively mitigate potential ethics risks down the line. [v] Operationalizing AI ethics must be more than a pipedream, more than a conversation. It’s time for organizations developing AI to ask, and answer, essential questions about ethical principles and processes. When you institute changes within your culture, you’re enabling the shift to an ethics-driven approach. For ultimate payoff, make the investment now in reframing the AI system design process to incentivize data scientists and developers to build systems built to your ethical principles. It’s time to move beyond ideals to action.

How to Embed AI Ethics into your Work Culture

AI Is The Future Of The Healthcare Industry

AI is not just a techy buzzword anymore. We regularly come across AI/ML technology in both personal and business use applications. Businesses are looking at AI for ways to increase revenue, improve operations, and to enhance productivity. If you are looking for readily available AI solutions that could accelerate your business, then dive straight in. We will be exploring some unique but probably underrated scenarios for introducing AI solutions in businesses. You will also discover how can help you leverage AI to achieve business goals. Let us start by understanding which areas of business can benefit from AI. Sentient.io What is The Democratization of AI? According to a recent , 77% of the organizations surveyed planned to either increase or retain their AI investments in spite of the global pandemic. Additionally, the report also talks about the “Democratization of AI”. This implies that businesses want everyone linked to the organization to experience the benefits of AI solutions. This includes customers, business partners, business executives, salespeople, assembly line workers, application developers, and IT operations professionals. How exactly can this sort of democratization be achieved when every aspect of the business functions differently? Gartner report In spite of the differences, there are many common practices across departments that are tedious and still carried out manually. For example, why do we still need someone to jot down meeting notes? While most businesses appreciate the value that AI solutions can provide, they are unaware of the common areas where automation can increase productivity. Fear of the unknown acts as a barrier to getting started. Also, acquiring the required expertise or knowledge to build and subsequently work with AI solutions is another challenge. Sentient.io can help businesses by providing a trusted source of expertise to get started on their AI journey. Let us find out the relation between the different business goals and available AI solutions. How To Transform Your Business With AI There are several ways in which AI can be integrated into your day-to-day operations. Let us look at how different AI solutions can be used to achieve specific business goals. : Your sales revenue largely depends on better customer understanding and improved customer service. AI solutions that analyze data from both new sources like CCTV and beacons, in combination with existing sources like POS can provide a deeper understanding of customers/users. This can help to influence customer spending habits and increase stickiness. AI solutions that suggest optimal product placements in retail stores or online product recommendations are results of such analysis. Increase revenue : To improve operations, you need to identify redundant or lengthy processes that can be optimized. Again, operational data analyses can be used to identify and gain insights into operational inefficiencies. These can be subsequently addressed using problem-specific AI solutions that eventually result in a reduction in unnecessary operational costs. Improve operations : AI solutions provide intelligent tools that can reduce effort, time, and financial investments by automating manual or expertise-heavy processes. They can be trained to deliver an equal or similar quality of output at a fraction of the initial resource investments. Enhance productivity Let us now look at a few common business scenarios that can be optimized with the use of AI. AI Applications That Address Your Mundane Problems We will now look in detail at five challenges that are common to many businesses and AI solutions that can be used to address these challenges. 1. Meeting notes The process of taking down meeting notes usually does not cover all spoken content and is prone to human errors. Without recorded texts of the meeting, it is impossible to run a textual search to find or reference discussions that happened in the meeting. The result is lost productivity (in recalling or listening through the recorded audio, if any) and errors (in an incorrect recall or misinterpretation due to imperfect memory). Voice AI solutions like from Sentient.io can process such recordings and transcribe them to textual meeting notes with high accuracy. This text may be further processed to add context, before sharing it with all participants and stakeholders. automatic speech recognition 2. Customer profiling In the retail industry, customer profiling helps businesses to understand the spending habits of customers and design targeted marketing campaigns aimed at a particular type of customer. Customer profiling is usually unidimensional, from a single data source that belies the multifaceted nature of customer behaviour. Such data may not be enough to provide a good understanding of customers and subsequently result in poorly informed business decisions. Data about customers may be available from other sources, which can help to build an accurate picture of the customer. For example, a retail chain may combine CCTV data with POS data to better profile their target customers at different locations. This would help to increase revenue while allowing for better management of inventories.  A Sentient.io in specific departments of the store can help to identify items that were seen but not purchased by customers. solution that can count the density of people 3. Podcast production The podcast production process requires an effective voice and clear-speaking capabilities, in addition to a lengthy process of metadata creation. You may also want the podcasts to be delivered in specific accents depending on the audience demographics. The cost of production is directly proportional to the human effort involved in most cases. However, AI-based services from Sentient.io can help you to not only edit but also create podcasts from scratch. Such services are available as affordable SaaS solutions. text to speech 4. Extraction of financial information Businesses may be required to extract financial information about individuals in situations where background checks are required. Extraction of such information manually is a laborious process because the information available is unstructured and distributed across different types of sources. The Sentient.io individual search combines data from different sources and provides it in a readable and manipulatable format thereby addressing the operational inefficiencies of the extraction process. solution 5. Video Editing Creating short-form videos from long-form videos is a laborious and time-consuming process that requires adequate skills and experience. There is a high cost attached if you engage internal or third-party resources to perform this task. An AI-based video processing software could provide the solution to this problem by analyzing the video content to identify key scenes in the video based on certain criteria and using these to build a shorter video.  This solution would eliminate the requirement for human effort and provide businesses with an affordable SaaS solution. How can Sentient.io help Sentient.io provides many diverse and valuable tools and solutions through their AI and Data microservices platform that can address varying business needs across industries. Sentient.io microservices can be easily integrated into your solutions and provide pre-built and pre-tested AI functionality in the following areas. 1. Computer Vision Object Detection People Counting People Recognition 2. Natural Language Processing Word-sense Disambiguation Entity Recognition (Pick out key entities from a sentence) 3. Voice AI Speech to Text (Automatic Speech Recognition) Text to Speech Additionally, it provides data services by collating information from trusted sources. These services can be used to perform background checks on individuals as well as businesses from the convenience of your office. Similar to AI microservices, these analytics services also provide easy integration features that can facilitate processes such as one-click loan review in banks and financial institutions. Thus, Sentient.io can help you address multiple business challenges through AI solutions that can be integrated into your business processes. Here we have presented a few scenarios, but there are many more. To learn about the many other ways in which your enterprise can benefit from AI solutions, download our full report available at https://www.sentient.io/ai-colab-lp?utm_campaign=Signups_Content_01_2021_Five_remarkable_AI_solutions_that_help_to_achieve_business_goals&utm_source=Hackernoon Previously published at https://brandzasia.com/five-remarkable-ai-solutions-that-help-to-achieve-business-goals/

ThirdAI

thirdai.com

11-50 employees

Since 2021

COMPANY RANKING

THIRDAI

EVERGREEN INDEX #10179

ThirdAI WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

ThirdAI

thirdai.com

11-50 employees

Since 2021

COMPANY RANKING

THIRDAI

EVERGREEN INDEX #10179

RELATED COMPANIESRANK

STORIES ABOUT THIRDAI

ThirdAI WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!