MetaNovas Biotech is a vanguard in the health and beauty industry, acclaimed for our strategic partnerships with illustrious global brands such as L'Oréal, Beiersdorf, Haleon, Takeda, and Unilever. Our forte lies in utilizing cutting-edge AI and advanced biotechnology to innovate and elevate the standards in cosmetics, medical aesthetics, and functional foods. Driven by a pioneering spirit, we are committed to unlocking groundbreaking active ingredients and formulating products that not only deliver visible results but also enhance overall well-being. As we continue to forge ahead, MetaNovas Biotech remains dedicated to blending scientific discovery with practical application, providing our partners with solutions that are both transformative and sustainable. Our collaborations are testament to our expertise and the trust placed in us by the leaders of the industry to bring to market the next generation of health and beauty products. Join us at MetaNovas Biotech as we shape the future of personal care and nutrition, creating a legacy of innovation and excellence.

MetaNovas

metanovas

Welcome Founders! See other interviews here. Hey Hackers, Metana has been nominated in HackerNoon's annual awards in Boston, United States. Startup of the Year Please vote for us here Read more about us below to understand why we deserve your vote. Meet Metana is a startup focused on providing education and boot camps for existing developers to become proficient in blockchain development. Their 4-month bootcamp program trains developers in technical skills and soft skills necessary for their careers. They also offer a job guarantee, promising to refund tuition if their graduates are not hired. Metana Metana's Web3 Solidity bootcamp has been running successfully with over 50 students and 10+ graduates hired in the industry. We are currently in our 12th iteration, and the program continues to grow in popularity. Our rigorous and hands-on approach to web3 education has been well-received by students, resulting in high satisfaction rates and successful job placements. Metana is not involved in any crypto-related activities or advice. My Role As the CEO and Founder of Metana, I am passionate about providing individuals with the skills they need to thrive in the rapidly-evolving tech industry. At Metana, we offer comprehensive bootcamps in blockchain development. equipping our students with the knowledge and hands-on experience necessary to succeed in these fields. In addition to our bootcamps, we also offer a JobCamp program to help our graduates further enhance their technical skills and land their dream jobs. How We're Improving the web3 Industry At Metana, we are disrupting and improving the web3 industry by offering advanced bootcamp programs in Solidity, Ethereum blockchain applications, De-fi, and smart contracts. Our programs provides Highly curated content Guided one on one mentorship with experts JobCamp, Job assistance that will help you with your first job and beyond. Community and collaboration with like minded people Hands on experience Expert designed roadmap Robust LMS Standing Out from The Crowd Metana stands out from other startups in the web3 and blockchain education space through its extremely rigorous curriculum, small cohort sizes, and hands-on emphasis. The program's focus on practice and problem-solving ensures students spend at least 80% of their learning hours coding or hacking. The remaining 20% is dedicated to theory, including understanding the fundamentals and minute details of how things work. Metana's curriculum goes beyond what is required by other courses, challenging students to write non-trivial smart contracts in assembly, break incorrectly used public signature cryptography, and recreate hacks that have drained applications of millions of dollars. Furthermore, the program limits each cohort to 10 students to ensure maximally effective learning outcomes and personalised attention from instructors. Our Predictions/Thoughts on the web3 Industry in 2023 As we move towards 2023, the Web3 industry is expected to witness significant growth and transformation. With decentralized networks gaining more traction, there will be increased demand for skilled Web3 developers and auditors. Companies will continue to invest heavily in blockchain technologies and De-fi protocols, driving innovation and creating new opportunities for developers. Metana's Web3 Solidity Bootcamp is well-positioned to equip learners with the necessary skills to excel in the Web3 industry. The curriculum covers a wide range of topics, including Ethereum Blockchain, De-fi, smart contracts, and more. With a project-based approach, learners can gain hands-on experience building real-world applications while receiving expert mentorship from seasoned instructors. As we move towards a more decentralized and private internet, the skills and knowledge provided by the Web3 Solidity Bootcamp will become increasingly valuable. Graduates of the program can expect to secure high-paying roles as Smart Contract Developers, Web3 Developers, or Smart Contract Auditors, with senior developers commanding salaries of up to $200k per year. What word defines the state of web3 in 2023? The word that defines the state of web3 in 2023 is "disruptive." The technology behind web3 will continue to shake up traditional industries, creating new business models and opportunities for entrepreneurs and developers alike. Web3 technologies such as blockchain, smart contracts, and decentralized finance ( ) will continue to evolve and become more mainstream, providing a more trustworthy and transparent way to conduct transactions and share data. DeFi Why we decided to participate in HackerNoon's Startup of the Year awards Metana decided to participate in HackerNoon's Startup of the Year awards because we are committed to providing the best web3 Solidity bootcamp program available. We believe that our program is unlike any other in terms of its collaboration, one on one mentorships, and emphasis on practical experience. We want to showcase our bootcamp program to the world, and the Startup of the Year awards provide an opportunity to do so. Being a participant in the Startup of the Year awards also aligns with our mission to bring blockchain education to the masses. The recognition from HackerNoon would help us broaden our reach and enable more people to learn about our bootcamp programs. It would also provide validation of our approach and the quality of our programs, which is invaluable in a crowded marketplace. We are honored to be part of the Startup of the Year awards and excited to showcase how we are disrupting and improving the web3 industry with our innovative approach to blockchain education. Final Thoughts In conclusion, Metana is committed to disrupting and improving the web3 industry, as well as the field of artificial intelligence (AI) and machine learning (ML), through our innovative approach to education. Our Web3 Solidity bootcamp program is designed to equip learners with the knowledge and skills needed to excel in the rapidly-evolving web3 industry. We believe that the future of the Internet lies in decentralization and the adoption of blockchain technology, and we are proud to be part of this movement. Additionally, we plan to bring fresh ideas and approaches to the AI/ML industry with our upcoming courses. Through our bootcamp programs, students receive personalized attention, hands-on experience, and expert mentorship, enabling them to become skilled blockchain developers, auditors, and AI/ML professionals. We are passionate about empowering people to leverage these transformative technologies to build a better and more equitable future. We look forward to continuing to lead the charge in web3, AI, and ML education, as we strive to disrupt and improve industries through our innovative and project-based approach to education. Metana has been nominated as Startup of the Year in Boston, MA, North America. Vote for us today!

Educating a New Era of Blockchain Developers - Interview with Startups of the Year Nominee, Metana

Yes, the NFT has burst, and for good reason. This particular bubble was caused by folks looking for the next project that could be a punk or ape, and speculating on the concept versus there being any real tangible value that drove the growth. bubble As such, it has come as no surprise that there has been somewhat of a collapse in those projects, not to mention the numerous rug-pulls which have stolen funds from speculators. The Inflation of the NFT Bubble With , the sudden growth in 2021 was unanticipated, but if you look at the history of NFTs, they have been around for a while. Crypto archaeologists are still arguing over what project first brought about the idea of non-fungibility to the blockchain. NFTs There were clearly some trigger events that kicked off the sudden in 2021, for example, Top Shots launching and artists who had been experimenting with the space suddenly found huge followings, such as XCOPY and Trevor Jones. growth NBA On the fringes, there was HEN kicking off a revolution in the way artists and collectors interacted, and stepping in to buy Beeple’s ‘First 1000 days.’ Metakovan These events contributed to a frenzy and general positivity in the crypto markets to the rapid expansion of NFT projects. But there was also talk that 99% of the projects that were springing up would collapse, as they were mere copycats with slight derivatives of successful projects such as the Crypto Punks or Bored Ape Yacht Club (BAYC). In essence, this is not a one-off. Any new concept will have excitement and rapid growth, for example, the sudden excitement around ChatGPT. And there will be a collapse as speculators realize that there is nothing tangible backing a lot of the ‘me toos,’ which then results in a retrace. This is normal market dynamics. The Plunge in Popularity There are several types of NFTs that have weathered the storm pretty well. Take physical tokens or the gold standard projects, such as Crypto Punks and BAYC, and the ‘real’ digital art space. If you look at the floor prices for Crypto Punks and BAYC, in ETH terms, the floors have not collapsed. With art, again, marketplaces on Tezos, SuperRare, and NiftyGateway are doing reasonably well with creators continuing to produce work and collectors participating. Physical asset-backed tokens are also generally shielded from this noise as they have tangible value and are not driven by speculation. What has not fared well are all the copycat projects which have added little to no value, driven by the concept of effectively fleecing the communities they claim to be building. A lot of the creators of these projects have realized how hard it is to build and simply walk away. So there has been a bursting of the bubble, but it’s a bursting of the bubble around valuable creators and projects – and this sets up the space well for the future. NFTs That Didn’t Survive the Hype A number of factors caused the downfall. For instance, the wider macroeconomic climate made people realize that speculating on the next cartoon animal isn’t the wisest decision in these markets. Another factor was the crypto winter, with prices of all tokens falling and folks sitting on major book losses. Why amplify the losses by participating in these speculative games? It’s too risky for everyone except the crypto diehards. Who remembers Kevin? There were some poor outputs from much-hyped projects, which has added to the general dismay around where the NFT market was heading. Without new buyers to offload useless Profile Picture NFTS (PFPs) to, it’s no surprise that prices have collapsed, and the last round of folks are left holding useless bags. The best signal is the rate of release of PFP projects. Towards the peak of any bubble, there is a rapid increase in copycats doing the same thing with little innovation, simply riding the bandwagon of free crypto money. All it takes is a trigger event – such as investors fleeing crypto into safe-haven assets – and suddenly, all that promised utility wasn’t worth the bytes used to store those NFTs. For the segments of the NFT market that have not collapsed, it’s clear that there is real value there. The NFT Comeback The technology behind NFTs can be used for solid use cases. Art – not generative projects – created by real artists will continue to thrive. A lot of the artists are expanding into other platforms to offer their work or doing it directly using the work of the Manifold team. Gold standard projects will continue to grow because there are sizable funds protecting them from failure at this stage. There’s excitement around blockchain gaming, despite the initial backlash to NFTs being associated with games, which is mainly due to a misunderstanding of how the technology can be used. NFT technology has the potential to improve many industries. We’ll see more innovative applications over the next couple of years, and there may even be a separation between crypto and NFTs. For example, users no longer need crypto to pay for NFTs. In terms of PFP projects, hopefully, folks do more due diligence on the competence of the teams behind such projects and the potential graphics that are produced to ensure those teams can really deliver. Now that we’ve exhausted every cartoon animal out there, hopefully, the next wave of projects look at more interesting solutions for collectibles that have intrinsic value.

Some NFT Projects Have Weathered the Storm but the Majority Failed

F1 Delta Time was a blockchain-based game developed by Animoca Brands that was offered to gamers, firstly in the beta form, between 2019-2022. The game went up in flames in March 2022, to much rejoicing from all NFT haters. But was this hate deserved? How much of a trainwreck was the game actually? This is my hypothetical post-mortem based on online evidence, with the aim to evaluate the value that was added to the game by implementing blockchain technologies. I haven’t played the game myself, and I honestly never cared about it, although I’ve been aware of it since 2019. At that time, I could not believe that a racing game could be built on blockchain. Now I am playing a lot of Hill Climb Racing 2, and I am completely convinced that you can build a racing game as good as Hill Climb Racing, on the blockchain. I urge Fingersoft to not take this as advice. How Exactly Does F1 Delta Time Work? From what is now possible to , F1 Delta Time was a browser racing game that utilized blockchain technology to assign exchange value to game assets. Racing is one of the most fast-paced genres by definition, and the only way blockchain could help was to support its economy. collect from YouTube participation in each race cost a relatively small amount of fungible tokens, akin to inserting a coin into the arcade machine. However, it included a certain element of skill, even if somewhat levelled down with expensive customization that followed a ‘pay-to-win’ trend. The game was ‘pay-to-play’: Initially, F1 Delta Time was built on Ethereum, which . To solve this problem, the game’s own fungible token, . can be used for participating in races and buying cars and car parts on open marketplaces based on Ethereum and Polygon. A part of the supply was airdropped directly to players, while the rest of the ‘monetary reserve’ was allocated to future prize pools and could be earned by winning races and lending out racing tracks to other players to compete. first became unusable in the summer of 2020 REVV, was introduced in September 2020 This token Surprisingly, as compared to many other ‘minor altcoins’, the price of the REVV token during its existence, although it naturally followed major cryptocurrency price trends, as well as major hype-generating events such as big sales in the first year of the game’s existence (Fig.1). has proven to be relatively stable The unique selling point of F1 Delta Time came from the official collaboration with Formula 1® and the value that was associated with the brand. This collaboration made F1 Delta Time a premium product and justified the high costs of its non-fungible tokens that represented Formula One cars and drivers. Several notable purchases have established F1 NFTs among on the already bubbling market. The first one happened in May 2019, long before the official launch of the game and even before the REVV altcoin was introduced. The buyer, then known under the name ‘ ’, , for the equivalent of $69 million from Beeple at Christie’s. the most expensive ones MetaKovan would later be revealed as Vignesh Sundaresan the same person who would later purchase the most expensive art NFT to date Back then, in 2019, MetaKovan paid the equivalent of for the unique race car in F1 Delta Time. The car had its own name ‘1-1-1’ and was all covered in digital diamonds. Notably, in his interview with Jon Jordan for the Blockchain Gaming World podcast, MetaKovan has described one more way to assign value to NFTs, to which we will return in our future cases: the value of a story. $111,000 in ETH The second, even more expensive, the sale happened in December 2020, three months after the game went public. , and he paid 9 million REVV (223,000 dollars at the time) for the token that represented a track segment for the Grand Prix of Monaco. As we already mentioned, the owner of a track segment could generate passive income by collecting a share of the racing fee when other players used this track for a race. Vignesh ‘MetaKovan’ Sundaresan was, again, the buyer There were s later, which broke the record set by MetaKovan, but we will use these two sales as most showing examples of the value of NFTs comes. other major sales of F1 NFT The question we are exploring here is how important the value of a brand is in the world of NFTs. F1 is a unique example of how this value went to zero overnight. Animoca Brands received a crushing blow to its reputation when of the 3-year licensing agreement with Formula 1 on March 15, 2022. they announced the end The game was shut down the next day, although its former players will still be able to swap their F1 Delta Time cars for other assets by July 31, 2022. The PR backlash resulted in at least a month of negative press from every gaming website beyond the crypto corner of the internet. Now, when the dust has settled, it may be that Animoca Brands is still going strong in terms of blockchain gaming - and even the assets that should be considered worthless without the power of a brand are still branded and up for sale after two months since the end of the licensing agreement. Even though the game itself can’t be played - only a line of code shows up in the browser where the playground for most expensive NFTs used to be - branded game NFTs are still available on OpenSea, available for purchase (see Fig. 3). Even the Monaco racing track that belongs to MetaKovan is still listed - to be precise, , now representing nothing but the loss of 9 million REVV, is still worth about $250,000 in May 2022. The image file still remains on the server, though, as well as all other digital images that represent deprecated game assets, and rare transactions still take place on OpenSea. the .png render of the track It may be that the value of a brand does not disappear with the end of licensing agreement. Even more, its role in the valuation of game assets may not have been as important as the critics suggest. What F1 cars and drivers have actually lost - in a rather painful way - is their utility value, that is, the ability to use assets productively in the game. For example, MetaKovan can’t generate passive income by lending out of the Monaco track segment. Of course, this .png will be deleted one day after the server clean-up, - but the same is true for many other unique and premium NFTs in many reputable projects that have, in Bondian terms, no time to die, although they may as well die another day. However, legendary and exclusive cars and drivers still retain potential exchange value: for instance, MetaKovan can still sell his 1-1-1 car on OpenSea. To make it almost actually immutable, the visualization of this car NFT on OpenSea is currently . the .png file an (unlisted) YouTube video Is F1 Really A Failure? It has been definitely painted as such in the media. However, if we look at the game statistics , it appears that the game’s popularity had been fading for many weeks already before the partnership was discontinued. available on DappRadar The number of daily players rarely exceeded 10 since September 2021 - for 6 months already - which would not be enough to continue a costly partnership with Formula 1. Of course, these are only the statistics that are publicly available at DappRadar, and they are already quite buggy, but there is no reason to suggest that the game was more popular than that. In the meantime, Animoca Brand launched its own racing game REVV Racing in August 2021, as the hype around F1 Delta Time was fading. It is very likely that the remaining players of F1 Delta Time have moved to that game even before the closure. In terms of player engagement, by far outpaced F1 Delta Time, although we do not have the data about how profitable it was for the developer or for an average player. REVV Racing As of May 2022, the collection of amounted to 167,000 items that belonged to 59,400 players. This is a success if we compare it to F1 Delta, which , and which belonged to 1,800 players, which is the cumulative number of all players during the lifetime of the game. REVV Racing cars on OpenSea only had less than 4,000 cars This large number of REVV Racing players can be explained by : everybody who had 8,000, 12,000, and 20,000 REVV tokens in their wallets would also get one, two, and three free new cars. Even though these were still considerable investments (8,000 REVV would amount to 1,137 USD at the time of airdrop, July 16, 2021), many free cars would go to the secondary market, and even more new cars would be minted and distributed later. From the viewpoint of an outsider, it looks as if the company moved from monetizing through ultra-wealthy ‘whales’ to a wider, poorer, but also more active and motivated player base. the airdrop rules Why communicate ‘deep regret’, then? The main problem with crypto games today is that, due to the rapid speed of the NFT scene, many of their creators don’t develop long term memory, so they never learn (or maybe, they are just not used to hiring the people who actually do the work of development and marketing, for the real money, on a long term basis). A similar event already happened in 2018 in CryptoKitties - the first popular blockchain-based game that also tried cooperating with sports brands at some point. In early May 2018, the company prepared to launch a marketing campaign CurryKitties, to celebrate their first NFT based on a living celebrity, the NBA player Stephen Curry. The deal was canceled at the last moment about the role of Stephen Curry in the promotional campaign. The basketball-themed kitties were already pre-programmed to be minted by smart contracts, and Stephen Curry was replaced by at the last minute. Although some fallout from the CurryKitties could be observed in social media, it was by far incomparable to the grilling that F1 Delta Time received from all kinds of media in spring 2022. due to a misunderstanding a generic character Swish Based on that, the devil’s advocate in me suggests that Animoca Brands should have communicated smugness, not regret, in their public announcement. Anything along the lines of “Formula 1 is not good enough for us” would please their core player base and provide less food for the trolls. Griefing could be easily redirected at the former partner, whose reputation is by no means squeaky clean. In the end, if the value of a brand does not convert into other types of value in this particular community of players, why keep this brand? How Important Is the F1 Disaster At All? For the company that has the word ‘Brands’ in its name, and describes its mission as bringing premium brands into the blockchain and NFT space, - yes, it is a big screw-up in terms of PR. However, their other game REVV Racing seems to be doing really well on the ‘play-to-earn’ scene, based on publicly available information. There’s no value in that brand, but the main point is that it perfectly communicates the message that sells right now: this is where you can earn minor altcoins by playing a simple racing game. As you can see, upon a very rough calculation, the price of REVV is well correlated with the number of active players. Moreover, as the new crypto winter is coming, we are about to see what will happen to the game if the price of REVV goes down. The number of active players after that will demonstrate how much of REVV’s value is utility value in the game, and how much is exchange value, boosted by speculation, lending, and borrowing, as in . the recent case of Axie Infinity Thank you for reading! We are living in a time when I don’t even want to expose my Ethereum wallet for possible donations because some of you might want to send me an exploding NFT. However, let me know if you want to see this research continued in a more academic manner. It will be even more fun to revisit this case after July 2022, when the migration of former F1 Delta Time players is finished. The statistics here are super basic and only used to illustrate very obvious trends. It would take much more effort to collect better data and produce objective conclusions, and I have to do other things right now. See you when I’m a bloody Doctor of Philosophy in Communications!

F1 Delta Time: The Value Of A Brand

NFTs have become a creative and commercial revolution. They have become a major topic of discussion among artists and entrepreneurs worldwide and have risen to an unimaginable value. As per recent estimates, the NFT market has surpassed in 2021, compared to just under $41 million in 2018! $40 billion NFTs have been sold for millions of dollars. And here is a look at the top 10 largest NFT sales that have happened in 2021. Top 10 Most Expensive NFTs on the Planet as of 2021: ‘The Merge’: Pak, $91.8M ‘Everydays - The First 5,000 Days’: Beeple, $69.3M ‘Human One’: Beeple, $28.98M ‘Cryptopunk #7523’: Larva Labs, $11.75M ‘Cryptopunk #4156’: Larva Labs, $10.35M ‘Cryptopunk #7804’: Larva Labs, $7.6M ‘Cryptopunk #3100’: Larva Labs, $7.51M ‘Right-click and Save As guy’, Xcopy, $7.09M ‘Ringers #109’: Art Blocks, $6.93M ‘Crossroad’: Beeple, $6.6M Deciphering NFTs: What are Non-Fungible Tokens (NFTs)? The term ‘NFT’ stands for a ‘Non-fungible token’. If an asset can be exchanged or substituted with a similar asset having the same value, it is ‘fungible’. For instance, the currency is a fungible asset. If you take ten $1 bills and exchange them for a single $10 bill, the value of your money will still be the same. But non-fungible assets are one of their kind in the Cryptosphere. They can’t be traded or substituted for something similar because nothing similar exists! For example, the original Mona Lisa painting is a unique piece of art. It cannot be exchanged for a Mona Lisa poster sold at a local gift shop. The Eiffel tower cannot be exchanged for its photograph taken by somebody, because only one of it exists. People take a unique piece of creative artwork and assign a digital token to it. This token acts as a certificate of ownership and authenticity. And it gets stored on a blockchain ledger. Whoever has that token becomes the owner of that property and its rights. It can be a traditional or digital painting, video games, sports memorabilia, songs, GIFs, tweets, a social media post, anything! Since such artworks are immensely rare and unique in their design, they cannot be duplicated. Thus, they have a very high value. Top 10 NFT Sales in 2021 Let us look at these famous NFT sales in more detail: Pak’s ‘The Merge’ Beeple held the record for selling the most expensive NFT of 2021 for most of the year. But from December 2 to December 4, 2021, the NFT artist Pak put ‘The Merge’ on sale via Nifty Gateway. The artwork was divided into numerous masses, each costing $575 and going up by __ __every six hours. $25 If a buyer already had a piece by Pak, the mass’s price would come down to $299. The piece collectively fetched $91.8M, with 312,686 pieces sold among 28,983 traders. Beeple’s ‘Everydays – The First 5,000 Days’ Mike Winkelmann, also known as Beeple, is one of the most prominent NFT artists today, with his own NFT marketplace as well, going by the name We.new. On 11 March 2021, sold this NFT through Christie’s online auction for $69.3M (42329.453 ETH). It was bought by Vignesh Sundaran, aka Metakovan, founder and financer of Metapurse. This particular sale was the most popular of the year and plunged NFTs into mainstream attention. Beeple’s ‘Human One’ The ‘Human One’ NFT is a life-sized 3D sculpture. It shows a human born in the Metaverse, wearing a silver-coloured space suit. The unique thing about this artwork is that it can change in form every day based on Beeple’s desires. He has promised that he will keep updating it during his lifetime. It was sold on November 9, 2021, to the Swiss entrepreneur Ryan Zurrer, for $28.98M. The sale happened through the 21st Century Evening Sale auction hosted by Christie’s.** Larva Labs’ ‘Cryptopunk #7523’ Larva Labs studio is a team of two Canadian software developers Matt Hall and John Watkinson. And on 10 June 2021, they sold this NFT for $11.75M at London’s Sotheby’s. The buyer was DraftKing’s largest shareholder, Shalom Meckenzie. It is also dubbed as the ‘Covid Alien’. It is the most expensive Cryptopunk NFT among Larva Labs’ Cryptopunk collection, containing 10,000 uniquely created characters. These punks are 24x24 pixel unique avatars in 8-bit style. And among cryptocurrency communities, people use them as status symbols due to their rare and exclusive nature. While some consider their Cryptopunk #9998 to be the most expensive NFT ever sold (for $532 Million!), its sale remains controversial. Larva Labs’ ‘Cryptopunk #4156’** ‘Cryptopunk #4156’ was sold for $10.35M (2500 ETH) to an unknown user on 9 December 2021. It is a rare Ape-styled punk. And there are only 24 of these in the entire 10,000 Cryptopunk collection.** ** Larva Labs’ ‘Cryptopunk #7804’ This NFT was sold for $7.56M (4200 ETH) on March 11, 2021, to someone called Peruggia on Twitter. It was one of the two major NFT sales that happened on the date. Larva Labs’ ‘Cryptopunk #3100’ Selling for a touch less than #7804, this Cryptopunk was sold for $7.51M. It was the second of the two major March 11 NFT sales. XCopy’s ‘Right-click and Save As Guy’ This artwork by XCopy sold for $7.09M through Superrare in November. It was bought by the rapper and crypto whale Snoop Dogg, under his NFT alias ‘Cozomo de Medici’. Art Blocks’ ‘Ringers #109’ Art Blocks is a platform that creates on-demand generative-styled art collectables. Among its collection of 99,000 NFTs, this one was sold for $6.93M to an unknown collector on Oct 2, 2021. Beeple’s ‘Crossroad’ Beeple struck again with this 3D animated 10-second short of his. This short featured the former US president Mr. Donald Trump and was first sold on Nifty Gateway. But it generated $6.6M via a secondary market sale on February 25, 2021, bought by an anonymous buyer. Conclusion With their demand rising exponentially around the globe, NFTs will revolutionize commercial transactions and ownership as we know it. Be it purchasing digital real-estate and owning unique creative assets or using NFTs in software licenses, event tickets, and club memberships, its use cases are endless. Coupled with the rapidly-growing Metaverse, NFTs might even change the face of the world.

Top 10 NFT Sales In 2021 That Broke All the Records

Brinc and Animoca Brands are working together to offer a program that advances the state of NFTs. 2021 is the year NFTs took over the world -- or at least the collective imagination of investors. August has been another record-breaking month for NFT sales. On August 18, NFT sales hit a record-breaking $897 million from the past 30-days according to . NonFungible The Launchpad Luna accelerator will identify and foster NFT innovation in the fields of culture, art, entertainment, media, gaming, streaming, collectibles, insurance, finance, and data management; DeFi and additional verticals will be added in the future. Just one doom-scroll through Twitter will show you that anyone who can get involved -- as a platform or as a creator -- is getting on board the trend. NFT Venture accelerators Brinc and digital entertainment, blockchain, and gamification company Animoca have joined forces to enable the flood of new NFT projects with their new NFT Accelerator Launchpad Luna. Early-stage projects and startups accepted into the acceleration program will receive practical training, a launchpad, and investments of up to USD$500,000 (or equivalent) in exchange for equity and/or tokens in the project. The new accelerator offers support from mentors and corporate partners working in the NFT space, including AppWorks, Blockparty, Dapper Labs (the company behind CryptoKitties, NBA Top Shot, and Flow Blockchain), EllioTrades, Featured by Binance, Gabby Dizon (co-founder of Yield Guild Games), Harmony (ONE), Hedera Hashgraph (HBAR), Mai Fujimoto (Miss Bitcoin), Mateen Soudagar (DCLBlogger), Metakovan (Metapurse), Mindfund, Sebastien Borget (co-founder of The Sandbox and chairman of the Blockchain Gamer Alliance), Virtually Human Studio (creators of ZED RUN), WhaleShark, and various others. NFTs have captured the collective imagination of an interesting cohort of creators, blockchain enthusiasts, and investors. The desire for new content is palpable -- and there is a need to help new NFT projects get off the ground. We spoke to Manav Gupta, founder and CEO of Brinc and Yat Siu, co-founder and chairman of Animoca Brands, to get their perspective on Launchpad Luna and what it is bringing to the NFT space. Manav Gupta, founder and CEO of Brinc (Left) and Yat Siu, co-founder and chairman of Animoca Brands (Right) What is the mission of Launchpad Luna? "Launchpad Luna is a collaboration between Animoca Brands and Brinc with the mission to invest in blockchain-focused startups that are shaping the future of the Web, driving digitalization, and enabling a more inclusive digital economy," Siu said. Who is your ideal partner for funding? "Ideal partners for funding are those that align with our mission to support an inclusive digital economy, who believe in the impact that NFTs can have on industries like art, music, culture, and finance, and who share our drive to support climate-conscious solutions," Gupta said. Do you mean to encourage creators by supporting art and NFTs? What do NFTs mean for creators? "We definitely want to encourage artists and various other creators. For the first batch of investments, the program will identify and foster NFT innovation in the fields of culture, art, entertainment, media, gaming, streaming, and collectibles. In terms of what NFTs mean to creators, NFTs are radically changing ownership, attribution, and remuneration. Primarily, NFTs represent true property rights for a digital asset, and the total control over an asset that comes with those rights: creators are truly the owners of their NFTs and can sell or trade them as they see fit.  NFTs also have the potential to allow creators to receive a royalty-like payment every time that the NFTs are traded by other people after the initial sale by the creator. NFTs also allow creators to provide much richer experiences to fans -- for example, you could have a single NFT that is a collectible, that can be used in one or more games (or similar products), and that has DeFi utility, all in one single NFT. A popular example of this is Animoca Brands’ REVV Motorsport ecosystem of racing games, where game assets are NFTs that have both collectible value as well as various utility. For example, the “1-1-1” is a famous Formula 1 NFT bought by Metakovan. The 1-1-1 has collectible value as the first-ever official Formula 1 NFT, it can be staked to generate rewards over time, and of course is a very powerful car that can be used to race in the game F1 Delta Time," Siu said. Can you share any NFT creators you are working within your accelerator program? "We work with a number of the top players in the industry. The accelerator has extensive support from various mentors and corporate partners working towards the advancement of the NFT space, including AppWorks, Blockparty, Dapper Labs (the company behind CryptoKitties, NBA Top Shot, and Flow Blockchain), EllioTrades, Featured by Binance, Gabby Dizon (co-founder of Yield Guild Games), Harmony (ONE), Hedera Hashgraph (HBAR), Mai Fujimoto (Miss Bitcoin), Mateen Soudagar (DCLBlogger), Metakovan (Metapurse), Mindfund, Sebastien Borget (co-founder of and chairman of the Blockchain Gamer Alliance), Virtually Human Studio (creators of ZED RUN), WhaleShark, and various others," Gupta said. The Sandbox For aspiring NFT creators and artists, what advice do you have? "See if you qualify for the Launchpad Luna program! This is where we really drill down into how NFTs should be made and used. In terms of general starting advice, all creators should carefully consider the smart contract that their NFT will use, and select the one that’s right for them. A unique (dedicated) smart contract is usually preferable to a cookie-cutter contract because it’s easier to establish ownership and to build in the features you need/want. If you are planning to sell your NFT as art, then make sure you know exactly where the actual image/video file component of your NFT will reside - solutions like the Interplanetary File System (IPFS) are better hosting solutions because they are more likely to preserve your image than a random file host. Finally, consider what utility you can give to your NFT in the rapidly growing world of blockchain products - this may require you to talk to different parties," Siu said. Is green consciousness important for the continued mainstreaming of crypto? Is it a prime consideration with Launchpad Luna? "Climate consciousness is one of the key requirements for Launchpad Luna. We want to shape the blockchain industry to embrace awareness of environmental concerns, especially climate change. Admission to the program is prioritized for climate-conscious projects that drive digitalization, that place emphasis on proof-of-stake protocols and sidechains instead of proof-of-work, and that have lower overall physical footprints. Accepted entrants will receive training in optimizing a blockchain business to minimize energy use and carbon emissions," Gupta said. It seems like art and collectibles are just the cherry on top of NFT use cases -- where do you see NFT use cases going and how will Animoca Brands help support the growing range of uses? "Art and collectibles represent only a tiny fraction of the potential applications of NFTs. We are especially excited about the applications for the world of gaming, an industry worth about $180 billion last year. Gamers already understand virtual goods and can appreciate the value of the property rights that come with NFTs. NFTs in games basically allow gamers to truly own their game assets instead of merely licensing them from game publishers. We expect this to be a major revolution in how gamers consume content and game publishers will need to adapt to a future scenario in which they no longer hold all the power," Siu said. Anything you'd like to add? "At Brinc and Animoca Brands we are all very excited about our joint Launchpad Luna program and we continue to strengthen the partnership between our companies. Brinc has made over 160 investments and is one of the world’s leading venture accelerators, while Animoca Brands has made over 75 investments in NFT-related companies, making Launchpad Luna an ideal combination of market-leading expertise in our respective fields," Gupta said. # # # Startups, creators, and projects that are looking for funding and support are invited to apply to the program . here Creators and artists who want to get involved with the program as mentors can inquire by emailing . partnerships@launchpadluna.com Cover image by metalandrew from Pixabay This article is part of The hosted by HackerNoon in partnership with Gaming Metaverse Writing Contest The Sandbox . Submit your today for your chance to win up to $2000. #gaming-metaverse story

NFT Accelerator Launchpad Luna is Helping Creators Reach New Heights

On 16th March, a piece of real-estate on the platform was purchased for $500,000! Somnium Space On 11th April, another plot of land on the platform was sold for $572,000! Decentraland Impressive, right? You might be wondering: Why would anyone buy non-existent land? In this article, I will explain where this trend originates and what drives investors to make such investments. What are Metaverses? Metaverses are the future of the online world. Metaverse is a shared virtual space based on the digital reproduction of our universe and is infused with a huge number of possibilities as well as the fulfillment of fantasies. The feeling of ‘presence’ in real-time is made possible in this space - social interaction (parties, conversations, walks with friends, virtual festivals, concerts, fashion exhibitions, sporting events, galleries, and more). All of the above contribute to developing a new economy – the NFT economy, which could grow into a real-world economy. To put it simply, a metaverse is known to everybody, i.e., Minecraft but on the blockchain. It’s the perfect ecosystem of the future. In regular games, it is up to developers to develop and bring in new elements. The user takes part in its development by creating games and events, building real-estate, adding their NFT objects, and paying in cryptocurrency. Some virtual worlds have their own cryptocurrencies; for instance, there is MANA on the Decentraland platform. Moreover, in a metaverse, unlike regular games, there is no script. The main purpose is to create social and economic interactions between users. Blockchain technology, which replaces physical and legal laws in the real world, ensures legitimacy and allows you to own your avatar, clothing, a virtual shop, or a piece of land in which such interactions occur. The key concept here is ‘ownership’, not just usage, as in ordinary multiplayer platforms. The economic dimension is now being actively integrated into such worlds. For example, a user can create an architectural object, then tokenize it and put it up for sale in various marketplaces. Later, this object will be bought by another user, after which it will be placed in their virtual space, and if so desired, NFT objects can be added to it. Do you remember Bepple's This NFT was purchased for 69 thousand U.S. dollars by Metakovan (Vignesh Sundaresan, a Singaporean investor). People might wonder why someone would buy an object of art that does not exist physically for such a sum? This is where metaverses come in. “Every day: the first 5,000 days”? Metakovan's ambitious plans include his intention to display his NFT in a virtual space. At the moment, the investor is working with engineers to design special gallery complexes so that everyone can visit the gallery simply by opening a web browser or using VR technologies. Just like Metakovan, you too can purchase a picture in the form of the NFT and hang it in your virtual house. Blockchain and NFT are the technologies that were lacking in Minecraft аs well as Roblox and Second Life. Technologies that are key to making metaverses are ironically realistic-looking. That is to say, they are a clear economic model that offers a sense of belonging for each user - a real property. The Metaverse Projects is the most developed metaverse and is based on smart contracts (LAND Contract, Estates Contract, Wearables, a Content Server) and assets. Its own cryptocurrency, called MANA, allows Decentraland to be an autonomous platform that subsidizes various operations and initiatives throughout the metaverse. Decentraland By the way, reported that during 2020, MANA Decentraland soared by 3500%+. A key feature of the platform is its focus on users' views since it is the community that votes for: policy updates, NFT-implemented contracts that will be globally active, land auctions, etc. On the grounds of this metaverse, users have already created many interesting objects, from castles to space bases. By the way, by the end of March 2021, the overall volume of Decentraland sales topped 50 million U.S. dollars. Coinbase is a virtual world on the Ethereum blockchain. Players can buy land on which they can build shops and art galleries. Editing tools, avatars, and text chat have already been installed. This metaverse is oriented toward art, where everyone can visit a number of different exhibitions. Cryptovoxels 'We create an open, social, virtual world; a world that has its own economy and its own currency.' This is the official motto of , a crowd-funding social world of virtual reality and virtual investments where people can buy land and dispose of it, building whatever they like: schools, cinemas, restaurants - the opportunities are unlimited. Somnium Space Why Buy Virtual Real-estate? So, we have come to the main question, which doesn't let many people sleep calmly at night. Really, why? Let's delve into this topic. As I have mentioned earlier, at the moment, Decentraland, Cryptovoxels, and Somnium Space are the leading platforms. Virtual real-estate prices in these metaverses continue to grow, and the value of some plots of land exceeds half a million U.S. dollars. Currently, each of these platforms has a few thousand landowners already. The rush to buy virtual real-estate can be compared to fighting for domain names when the Internet was just born. According to CEX.io data, the general capitalization of the NFT market in 2021 was $8 billion U.S. dollars, which exceeds the indices of 2019 by 56 times. At the beginning of 2021, the demand for NFT increased, and crypto-asset sales soared. More and more people are coming into the sphere, creating a competitive environment. For this reason, land plots in the central areas of virtual worlds will be in even greater demand and therefore need to be acquired as soon as possible. Virtual spaces are mainly real-estate that capture people's interest, shape the demand, and focus the attention of users. This attention can be monetized in different ways. Statistics show that virtual real-estate is booming now. Consider the metaverse Decentraland: in January, the virtual world had 111 buyers; in February, this figure increased to 184; and in March, there were already 334 buyers on the platform, with monthly land sales exceeding $4 million, having started at $246,000 two months earlier. Returning to the above-mentioned increasing demand and society’s focus on virtual land, I will give the example of an investor, Whale Shark, who collects NFTs. Starting in 2018, this investor bought the real-estate property in sandpits, with the total amount of purchases worth 400 ETH. Since then, the price of each land plot has risen from $60,000 to $400,000. Good investment, isn't it? By the way, some investors have no plans to return to real investments, keeping most of their savings in crypto assets (for example, the Australia-based investor Matin Saudagar, presenter of the NFT blog). The virtual world and NFT take society to the next level. Before long, humanity will be spending most of its time visiting galleries and events, participating in races, and building houses in virtual reality. Digital art, based on smart contracts, will contribute to these spaces' formation, development, and appearance. Land within metaverses will draw more attention from the general public, fuelling a passion for a new way of investing, i.e., buying virtual real estate. This trend is already gaining momentum!

NFTs, Metaverses and the Virtual Real-estate

The ArtCollecting.info team is pleased to announce the launch of a new NFT platform. This landmark event will be a new step towards the digital transformation of the art market. The release of the beta version is scheduled for June 5, 2021, the alpha version (final) - for July 5, 2021. The platform's peculiarity is the focus on a new generation of collectors involved in it projects. 95% of users who received an invitation to ArtCollecting.Info are the founders of DAO funds and owners of crypto assets: bitcoins, ethereums, Dogecoin, Tezos. Estonia will become the jurisdiction of the NFT platform. The digital transformation in industry and the financial sector has already brought about a shift in social paradigms and cultural codes. At the same time, large-scale implementation of cyber-physical systems and digitalization of industry are unthinkable without transformations in the humanitarian spheres. The February NFT revolution in the art market was a continuation of the industrial revolution, which is developing not linearly, but exponentially. The first collectors of digital art presented on NFT platforms were entrepreneurs involved in blockchain projects. In response to their demand, a new NFT platform for the ArtCollecting.Info team is being created. Its feature is integration with esports and betting platforms, which will make NFT auctions even more exciting and increase the demand for tokens. The new NFT platform will run on the Tezos blockchain domain Via Marina Nadeeva, founder of ArtCollecting.info: "Our team became a guide from the art world to the world of technology back in 2016. A few years before the lockdown, we talked about the role of e-commerce. In 2017, we carried out the first experiments with art tokenization." "In 2018 - 2019, more than 100 events were organized aimed at informing the art community about the possibilities of blockchain technology. We acted as partners of conferences in Davos, Basel, Miami. We held our own in Moscow (Unblock Art 2019). In January 2020, we presented a service for fixing copyright to an intellectual property object on the ArtCollecting.info platform using blockchain technology." "The opportunity to sell non-fungible tokens on the new NFT platform will be a new step towards the development of the art business. The first users of the platform will be artists who are not represented in galleries, but represented in the Metakovan collection. We reached them while surfing social networks - Instagram, Behance, Flickr, Pinterest." "Our NFT platform will become a platform for them to reach a new generation of collectors. There are already more than 160 of them in the closed Telegram group, created on the eve of the launch of the platform. In another Telegram group, New Art Market, more than 180 artists are discussing new strategies and prospects for the art market."

Art Collecting Launches NFT Platform on Tezos Blockchain Domain

Since FarmVille, gamers have been jumping from one farming game to the next in pursuit of the highest rewards. Digital tokens called non-fungible tokens (NFTs) have become a game-changer by giving gamers ownership rights to the in-game digital assets on their farm. That’s the allure of a new MMOG game called ’ whose farm animal NFTs are selling for an on the Chinese NFT marketplace Treasureland. The price of an MNA NFT for Berta the cow or Helga the sheep may seem exorbitant but it could be the best investment these gamers ever make in the blockchain. ‘My Neighbor Alice average price of 1,880 USD The social simulation game was designed to introduce gamers to blockchain gaming and the collectible digital tokens behind a global craze in gaming. NFTs turn digital assets into verifiable, tradable tokens that allow gamers to own their cherished digital goods. What they win, create, and buy in one gaming world they can now play, trade, and sell in another. This tamper-proof evidence of ownership in a collectibles market devalued by forgeries is behind a global boom in the trading of NFT-backed digital assets. The Chinese gamers are part of a surge in hobbyists during the COVID pandemic contributing to for newly tokenized collectibles such as sports trading cards. record high prices The throngs of gamers in China and Japan, the two largest online gaming markets, entering the NFT marketplace are pushing blockchain gaming into the mainstream. The infusion of NFTs into Asia’s gaming and consumer culture can be seen from China’s luxury goods market to Japan’s manga culture. The Rising Sun on Global Blockchain Gaming Readying for an onslaught of Asian gamers into blockchain gaming, more Asian crypto businesses are setting up shop in The Sandbox gaming metaverse. Most recently, leading Japanese crypto exchange Coincheck acquired a large number of LANDs in the form of ERC721 non-fungible tokens for its NFT marketplace. is the biggest development in the Japanese NFT market since J-Pop girl band top charter SKE48 instantly sold out its first set of NFTs in the fall. Japan’s beloved J-Pop bands, through the issuance of NFTs and promotion of crypto brands, have been one of several key influencers on gamers in the NFT market. CoinCheck’s NFT marketplace Asian investors, who have become the largest backers of NFT art, are another. Singapore-based investor Vignesh Sundaresan, known as Metakovan, recently for the first crypto art auctioned at Christie’s, Beeple’s ‘EVERYDAYS: THE FIRST 5000 DAYS.’ The runner-up was China’s Justin Sun of Tron and Rainberry (formerly BitTorrent). paid $69.3 million The Emerging Asian NFT Art Scene This year, crypto art will become fully immersed in Asian culture, as open and launch. The comics are being directed by none other than the creative director of Animoca Brands, the parent of The Sandbox, Kevin Abosch. Kevin, whose crypto art has sold for $1 million, is behind blockchain game F1 Delta Time, where tokenized Pirelli tires and engine blocks sell alongside $75k digital racing car collectibles as top-selling gaming NFTs. crypto art exhibits Manga-inspired crypto comic books Perhaps the ultimate status sign that NFTs have arrived in Asia is their presence in the Chinese luxury goods market. A pair of luxury sneakers is becoming a must-have item among some gamers. New Chinese digital brand is selling NFT-tokenized sneakers ranging in price from $1,500–$10,000, with one FEWO shoe by FEWOCIOUS recently . Atari sneakers with a holographic theme are currently best-selling NFTs. RTFKT fetching $28,000 Asia artists, like their counterparts elsewhere, see the opportunity to sell art and find jobs in the crypto economy. Over are listed on China’s BCAEX digital art gallery on the blockchain, which has sold close to half a million dollars in the crypto-collectibles. Meanwhile, Asian voxel artists are quickly making a name for themselves in NFT game art. of The Sandbox's most recent VoxEdit Celebrity Contest were from Japan. Many of them will find jobs creating digital assets for blockchain games. 10,000 tokenized artworks Two of the three winners Winning entry of voxel artist @tak_akkiy to The Sandbox VoxEdit Celebrity Contest Some of these gaming NFTs already have market value before the game releases, such as the charming avatars and fantastical creatures they fight in the ‘Hack the Tao’ game created by Italian artists Sergio Scalet and Nadia Squarci — a project incubated by . The award-winning ‘Hack the Tao’ crypto art has been exhibited globally. The Sandbox's Game Makers Fund Celebrity-backed brands are valued the highest. Digital game collectibles based on some of the more than 60 major IP brands including those of the Smurfs and Care Bears in The Sandbox already sell at a premium. Sandbox theme park owner to bring NFTs for gaming, music, and more to the Atari Metaverse. Rappers Ja Rule and Soulja, football great Rob Gronkowski, and techno-pop artist Grimes are some of the celebrities issuing NFTs, which could end up in games. Atari has partnered with Bondly The Global NFT Gaming Economy With digital wallets stashed full of NFTs, gamers can leave closed gaming worlds and participate in the global economy. Blockchain gaming shifts economic power from gaming owners to the players. The value of NFTs became material for many unemployed workers in developing countries during the COVID pandemic. In The Philippines, an ‘ ’ craze took hold as word spread that players were supplementing and even replacing their salary playing blockchain games. ‘Axi Infinity’ is the number one blockchain game on Ethereum, with close to 24,000 players. Axie Infinity , a community of blockchain gamers from The Philippines emerged to invest in NFT businesses and assets. The guild has set up its headquarters at its Sandbox estate, a recently acquired 144-plot estate, which is in addition to a 6x6 Sandbox estate and 88-plot Savanna estate in ‘Axie Infinity’. In addition to play2earn earnings potential, the guild expects members to benefit from job creation for guild members in open gaming economies. Yield Guild Games Because NFTs can be used across games on interoperable blockchains, the Chinese virtual farmers and Filipino Axie collectors could trade their digital items on other gaming platforms anywhere in the world. The NFT gaming world is rapidly expanding. Playtoearn lists 219 blockchain games with close to 10,000 NFTs already listed — a drop in the NFT bucket. The Sandbox has more than two dozen new games ready to launch in its metaverse. Each digital asset in a game — from a castle or dragon to a mushroom — can be turned into an NFT with the . VoxEdit digital asset and NFT maker

Farm Animal NFTs are Selling for an Average Price of 1,880 USD 

This is part 1 of my ISIC cancer classification series. You can find part 2 here . (Pictured Above: A malignant lesion from the ISIC dataset) Computer vision based melanoma diagnosis has been a side project of mine on and off for almost 2 years now, so I plan on making this the first of a short series of posts on the topic. This post is intended as a quick/informative read for those with basic experience looking for an introduction to the ISIC problem, and those just getting out of their first or second machine learning/data mining course who’d like a simple problem to get their hands dirty with. machine learning Tools for early diagnosis of different diseases are a major reason machine learning has a lot of people excited today. The process for these innovations is a long one: Labeled datasets need built, engineers and data scientists need trained, and each problem comes with its own set of edge cases that often make building robust classifiers very tricky (even for the experts). Here I’m going to focus on building a classifier. First I’ll show you a simple classifier, then talk about how we can measure its success, and how we can improve it. The early diagnosis challenge I’ll be exploring here is called the challenge. Here’s an excerpt of their problem statement: ISIC The International Skin Imaging Collaboration: Melanoma Project is an academia and industry partnership designed to facilitate the application of digital skin imaging to help reduce melanoma mortality. When recognized and treated in its earliest stages, melanoma is readily curable. Digital images of skin lesions can be used to educate professionals and the public in melanoma recognition as well as directly aid in the diagnosis of melanoma through teledermatology, clinical decision support, and automated diagnosis. I thought I would share some of the details for a simple classification system for this problem that can be built from only the things implemented in an early machine learning class. No or bleeding-edge research papers here; the goal of this post is only to build and analyze a classifier with modest performance that demonstrates how to judge your own entry into the ISIC challenge. The code is all on , so much of my work will be reusable by anyone interested in the problem. deep learning Github We’ll be using PCA (Principal Components Analysis) to reduce the dimensionality of the dataset, once the dimensionality is smaller we’ll train a random forest classifier on it. Then we’ll do some analysis on the classifier to see how it could be improved, and gather some important metrics for the ISIC challenge such as AUC (Area Under Curve), ROC curve, and confusion matrix. If you’re unfamiliar with any of these terms or why don’t worry, I’ll briefly overview them in their corresponding sections and provide references for further study. The Dataset I won’t dive deep into the details of the dataset, as the ISIC , but in this post we’ll focus on the binary classification portion of the challenge, and not the lesion segmentation, or dermoscopic feature extraction. In the binary classification challenge you’re asked to differentiate melanoma images from seborrheic keratosis and benign tumors, as well as seborrheic keratosis images from melanoma and benign tumors. I’ll focus on just the melanoma vs. others portion, as any work done that way can easily be swapped out to accomplish the seborrheic keratosis part of the challenge. It’s also important to know that the distribution of the dataset is 374 melanoma images, 254 seborrheic keratosis, and another 1372 benign images (also called nevus). More on how this skewed distribution will affect how we judge our classifier later. explains it all Dimensionality Reduction with PCA In image classification tasks individual pixels are your features, so dimensionality reduction is key. A typical 500x500 RGB image has 750,000 (500*500*3) features, which is intractable for any dataset numbering in the thousands, so we’ll need to employ a dimensionality reduction technique. Principal components analysis is what we’ll use to reduce the dimensionality. PCA will find the vectors accounting for the maximum variance in our data set, and we’ll use those vectors as our features (instead of the large number of individual pixels). You can pick the number of vectors you want to use for your dataset, and in our case we’ll use 20. If you’re new to machine learning, and have some undergraduate linear algebra skills, implementing PCA is a good exercise to learn the specifics. But if you don’t have a few hours to spare, think of PCA as an operation that will take us from a matrix to an matrix while retaining most of the information in our dataset (in our 20-feature case 80% of the variance is retained). At the cost of a longer preprocessing time, we could always increase the number of features to account for more of the variance of our data until we have features. MxN Mx20 N def reduce_dimensionality(dataset): data_with_class = np.load(dataset)
    data_no_class, y = extract_features_and_class(data_with_class)

    # Will Seg-fault regular PCA due to dataset size
    # Somewhat arbitrary batch size here. 
    pca = IncrementalPCA(n_components=N_COMPS)
    num_batches = int(math.ceil(y.shape[ ] / float(BATCH_SIZE))) i xrange(num_batches):
        batch = get_next_batch(data_no_class, i, BATCH_SIZE) 
        pca.partial_fit(batch)

    reduced_data = None i xrange(num_batches):
        batch = get_next_batch(data_no_class, i, BATCH_SIZE) 
        transformed_chunk = pca.transform(batch) reduced_data == None:
            reduced_data = transformed_chunk :
            reduced_data = np.vstack((reduced_data, transformed_chunk))
    reduced_data_with_class = np.hstack((reduced_data,y)) reduced_data_with_class
 
def get_next_batch(data, i, batch_size): (i + ) * batch_size > data.shape[ ]: data[i * batch_size:data.shape[ ], :] data[i * batch_size:(i + ) * batch_size, :]

def extract_features_and_class(data_with_class): y = data_with_class[:, ]
    # Reshape into column vector instead row
    y_col = y.reshape(y.size, )
    n_columns = data_with_class.shape[ ] - data_no_class = data_with_class[:, :n_columns] data_no_class, y_col "" " Reduces the dimensionality of dataset (assuming it stored using numpy) " "" with 0 for in for in if else return "" " Returns the ith batch of size batch_size from data. 
      If  (i + 1) * batch_size goes past the size of the data, 
      this just returns the remaining rows" "" if 1 0 return 0 return 1 "" " Seperates the features from the label. 
      Assumes the labels the final column of data_with_class. " "" -1 of 1 1 1 0 return The above is a little snippet of the PCA code I used in this classifier. was necessary here because the images are somewhat large; it adds a bit to the size/complexity of the code, but on the ISIC dataset the code above can still be completed on my laptop in 20 or so minutes (your mileage may vary; I’m running a Macbook with a 2.9gHz i9, and 16 gigs of RAM). It’s also worth noting that the images needed resizing in order to be uniform for the input into PCA. Incremental PCA Now that our dataset is smaller, let’s train a classifier on it! Classification For this I went with a tried-and-true classic: a random forest. The gist of this classifier is that some number of decision trees are built, each one trained using a random subset of the feature space. Then (as in other ensemble classifiers) the individual trees are used to vote on the classes of new examples. I really only picked the random forest because it tends to be a strong out-of-the-box classifier that trains quickly. The top papers for the ISIC challenge tend to utilize convolutional neural networks; these employ convolution and pooling operations on images in order to reduce their dimensionality before running the images into a plain neural network. Some papers even ensemble convolutional networks! I hope to release some details on how to do this effectively for the ISIC challenge in another blog post :) I chose not to include the code for the classification in this post as it’s pretty trivial (just a plug-and-play sklearn classifier). You can find the classification file on the Github page . here Analyzing Our Results The ISIC classification competition score is based on your classifier’s AUC (Area Under Curve) for its ROC (Receiver Operator Characteristic) curve. If you’re unfamiliar with these terms, the ROC curve is a plot of the true positive rate and false positive rate of our classifier at different classification thresholds. The ROC is handy because it let’s us see how much sensitivity we’ll sacrifice for a higher specificity. In melanoma diagnosis a low false-positive rate (AKA a high specificity) is key because any melanoma detection indicates the need for an invasive biopsy of the lesion. By changing the classification threshold, we can always give up sensitivity for specificity or vice versa; the ROC graph just shows us the trade off. AUC is the primary metric that your submission for the ISIC challenge will be judged upon. AUC is a crucial metric for imbalanced datasets because a high accuracy is often only indicative of a classifiers ability to guess the dominant label. Using the ROC along with its AUC we can get around any misleading accuracy percentages. The ISIC dataset you’ll download has far fewer melanoma examples than seborrheic keratosis, and nevus. Only about 20% of the default ISIC dataset is malignant, 374 images total. The skewed distribution has a big impact on how we judge our classifier, and how we train it. First of all, 80% accuracy would not be impressive on the initial dataset because we could get it by simply guessing “benign” everywhere. Any classifier that does this will end up with a low specificity and appear to be “guessing” on the ROC curve (a linear line on an ROC curve implies random guessing, which hurts AUC). Another problem with the imbalanced dataset is that your classifier will usually be built to minimize some loss function, so the imbalanced dataset will “trick” the classifier into only guessing the dominant class; even if it could make predictions based on something other than the dominant class’ prior probability. To remedy the skewed distribution we’ll undersample the dataset. In the case of the Stanford paper I mention below, I saw in their taxonomy that their percentage of benign lesions was 60%, so I went ahead and under-sampled until I had a 60/40 distribution. The under-sampling was random, AKA I just threw out random benign images (which isn’t the best approach), but more on that later. The undersampling dramatically improves the AUC for the PCA to random forest model we built. Before we jump into our classifier’s accuracies, sensitivities etc. we need to establish what a good score is. We’ll judge our classifier by comparing it to how dermatologists score today. Here are some ROC curves from a popular paper that came out this year titled . Nature “Dermatologist-level classification of skin cancer with deep neural networks” (The melanoma vs. all ROC curve for dermatologists in the paper mentioned above. The orientation is different than I’m personally used to because paper actually plots the sensitivity (true positive rate) vs the specificity (true negative rate). Most ROC curves plot the sensitivity on the y-axis and the false-positive rate on the x-axis, but this representation is equivalent. The difference between the two graphs is the way the photo was taken: dermoscopy images are taken with a special device called a dermoscope, and the others were taken in a normal clinical setting.) ) (The ROC curve for our classifier for each fold of our cross validation. Note that this is for both dermoscopy and non-dermoscopy images. I decided to flip this curve to match the one provided in the Stanford paper for easier comparison Look above, and you’ll be shocked to see that we didn’t do as well as a team of Stanford Ph.Ds… However, our graph isn’t bad for a first attempt. Note that it’s also likely that our classifier would perform worse in the real world due to the small size of the dataset (unlike the Stanford paper mentioned above which had two orders of magnitude more data). After doing that and running the classifier I got a mean accuracy of 75%, mean sensitivity of 73%, and mean precision of 71% across the 7 folds for the malignant detection task, and because we were doing a 60/40 split this does seem to indicate that our classifier is learning something! , dermatologists tend to get 60% accuracy unaided by dermoscopic imaging, and 75%-84% accuracy when aided by imaging techniques. So our classifier is almost as accurate as a dermatologist with the assistance of their dermatoscope, and likely better than the blind eye (assuming the dataset used to get that accuracy has a similar distribution as ours). According to the ISIC (The confusion matrix for one of the K-folds trials. The 60/40 imbalance in the dataset likely contributes to the higher misclassification rate for melanomas.) Our AUC for melanoma vs all is on average 80%. Top papers tend to have in the low 90s for the melanoma vs all AUC (and in the high 90s for other AUCs), so we still have a long way to go. Improving our Classifier Adding more data from the archive The ISIC provides an that using a REST API. Getting more malignant data this way so you don’t need to undersample is probably the easiest way to increase the performance of the classifier you just read about. archive of additional data you can query Better Sampling Techinques A core issue with the dataset is the underrepresentation of melanomas. If we train the classifier using a dataset comprised of 80% non-malignant data our classifier is heavily biased, resulting in very poor performance (you can do this experiment yourself quite easily, the AUC drops to about 65% as the sensitivity of the classifier becomes weaker). To mitigate the skewed distribution we’re currently undersampling the benign images, however a better approach may be to oversample the malignant examples with some noise. There are some pretty sophisticated oversampling techniques out there, like , which creates its own “synthetic” examples instead of simply duplicating the existing minority class. SMOTE Segmentation / Thresholding This is actually another totally separate subcategory within the ISIC challenge, but we can use it to crop the image to only sections where the lesion exist, and then train on those areas. This may improve classification accuracy by removing background “noise”. Choosing the right segmentation technique is tricky though, as many will fail and white-out the whole image. I actually include some of this code on the Github repository, and if you want to see that let me know. I’d be happy to make another post, or guide you through it. Utilizing associated metadata The ISIC also provides age, and sex information on the images, seeing if those affect the AUC for your classifier is low hanging fruit. In papers I’ve read, I’ve seen these features have an effect on the performance of convolutional neural networks in seborrheic keratosis classification, but not always in melanomas. Final Remarks If you’ve made it this far I hope it was helpful! You can find part 2 . here Originally I intended to use convolutional networks in this post, but I realized that with the analysis portion the post ended up being too long-winded (I worked a little to keep this around 2000 words), so I built a simple classifier so the guide could mostly focus on the analysis. If you find any mistakes definitely let me know! Also, the code was written in a somewhat hurried fashion, so if you’d like to reuse the code yourself as an exercise, but you’ve run into trouble, just message me and I can guide you through it. Thanks again :) : Minor grammatical mistake, added the second tutorial Edits

Machine Learning for ISIC Skin Cancer Classification Challenge

This article is presented to you by the Immortality Foundation. Aging is not an absolute, its a shapable relative. A chemical process in our cells, a genetically codified principle. But with the recent rapid biotechnological development, for the first time in history, we have the capabilities and the resources to determine our own fate. Unlocking the ability to attain indefinite lifespans is within reach, it would be fatuous not to stretch out and take it. Bright heads around the world are already working right now to defy Aging, but their number is small and there is so much more that can be done. Whether the defeat of Aging occurs in 20 years or 50 years is largely a question of funding. The best way to accelerate this process is by mobilizing those who desire the option of a longer and healthier life into a cohesive social force — crowdfunding relevant research and encouraging for its benefits to society. Genetic modifications in other species have already achieved great lifespan extensions, as seen for Caenorhabditis elegans (roundword) — by 10 times, Drosophila melanogaster (fruit fly) lifespan has been doubled and genetic mutations in mice can increase maximum lifespan to 1,5 times normal. Biological immortality will not only save up to 100'000 lives per day, but also spare Millions of aging people years of suffering, as well as give society a never seen before economy-boost. All funding will be used for three purposes, firstly, to setup the Human Immortality Foundation where research in areas mentioned in organisms and cellular Levels will go underway. Second, to create and strengthen international collaboration of scientists institutions working on longevity, and at last to encourage and support more scientists to get active in the field. The amount and speed of the desired research that can and will be done is completely dependent on the amount of funding acquired Right now our governments and public institutions are spending billions of dollars on research of ARD’s (age-related-diseases). This undertaking used to be very valuable, but as it is with science, when new discoveries are made and new knowledge comes to light, a shift of focus has to happen. In the recent past we just didn’t know better than to do research on Alzheimer, certain types of cancer, etc., but now we know that there is one root-cause to all of them: Aging. That means that there is a huge misdirected investment happening right now, which in the end will only be profitable for the big pharma companies (because they can continue selling you drugs for each condition). Since the major science community is only doing research on ARD’s, they are trying to fight the symptoms of ageing, its by-products. A fight against symptoms is always a losing battle in the long run, and anyway, what do you think is easier to tackle, 7 symptom problems or 1 core problem? THE ISSUE OF AGING Every day, 100'000 people are dying because of old age, most of them spend years suffering from age related conditions and the majority of healthcare goes to people over 60 years old. Aging is the single biggest risk factor of virtually every significant human disease. There is a reason why “diseases” of aging such as Cancer, Alzheimer’s, and Parkinson’s Disease are almost only showing up in older individuals. After living as a human for 50–60 years your risk of getting any of these are spiking dramatically. That’s because things are changing in your body. The sale of putative anti-aging products such as nutrition, skin care, hormone replacements, vitamins, supplements and herbs are massive, with the US market alone generating about $50 billion of revenue each year. Most of these do not even extend human life significantly or at all. If these same people were to support the research in anti-aging/reverse-aging instead with similar numbers, we would have access to longevity technologies exceedingly fast. The matter of rejuvenation is both interdisciplinary and multidisciplinary in nature, and we need the benefit of scale to be able to tackle many issues. The most important thing is to build a very good international infrastructure and create networks for the success to be flourishing. The field is enormously complex and this will be one of the first steps to achieve our goal and providing a network for scientists, labs and researchers around the world. Governments are not yet willing to acknowledge the potential and importance of such new medicinal improvements. But the clock is ticking and there’s no time to wait and rely on them, It’s time to step up now and get going, eventually they will jump on board, once they understand the magnitude of this. The problems of the world cannot possibly be solved by skeptics or cynics whose horizons are limited by the obvious realities. We need men who can dream of things that never were. — John F. Kennedy HOW CAN WE SOLVE THIS? Medical research tends to tackle age related diseases (ARD’s) separately. After all, the illnesses are distinct: cancer arises from mutated DNA; heart disease from clogged up blood vessels and dementia from damaged brain cells. The biological processes that underpin the pathologies vary vastly and each needs specific treatment. In the past, it may have been a good idea trying to treat each condition separately, because we didn’t know any better. It is fair to say that the greatest driver of disease in old age is old age itself and this is worked on way too little. Most of it happens through the accumulation of damage in our cells, nonetheless it can all be traced back to a decline in our cells ability to keep homeostasis (cell maintenance), which leads to the accumulation of damage in our bodies, ultimately making us sick. Our genes are responsible for cell maintenance, or moreover their own regulation. To put it simply, we need to instruct our cell’s information to keep the cell maintenance at a higher level than it currently operates at. Our cell maintenance is already pretty good, but it’s nowhere near 100% efficiency and there are other organisms out there which do that job much better, and we can learn from them. The most crucial element at this moment is to eradicate the basic mechanisms underlying these specific gene regulations through basic research. That’s the only way to get rid of the aging problem and all the problems that come with it for good. With such new acquired knowledge we can develop ant-aging therapies. They could then be put into effect through gene therapy, although medication may also be an option. The people who are crazy enough to think they can change the world, are the ones who do. — Steve Jobs WHAT DOES IT TAKE TO TREAT AGING? There are two approaches how it can be done. 1. Clean up the mess. It is clear that aging is the accumulation of damage in our cells and body, being embodied by what we call age-related diseases. One approach is to fight these symptoms of aging we all are familiar with. There are seven variables that lead to aging and there is research going on into every one of them right now (some more than others). Loss and atrophy or degeneration of cells Accumulation of cells that are not wanted Mutations in chromosomes and mitochondria The accumulation of “junk” within and outside the cell Cross-links in proteins outside the cell Once we can fix all seven of these things, we are done abolishing physical aging. With this approach it will be necessary to periodically remove these accumulated damages from our bodies, but that will be a small price to pay for near-immortality. 2. Keep it clean. Those symptoms of aging have all one common root cause, an inefficient cell-maintenance (homeostasis). There is strong evidence that aging and longevity are controlled by a multiplicity of molecular and cellular signaling events that interface with environmental factors to maintain cellular homeostasis. A decline of our cells capacity to keep high-efficient homeostasis leads to dysfunction of specific cell types and tissues, rendering the organism susceptible to a range of chronic diseases. Such decline is only possible through an intrinsic over-time alteration of our gene-expression. The only way to stop this from happening is, to determine which specific genes are involved in this process, so that their expression-setting can be reprogrammed, ultimately making you stay and look young without a time-limit. This method would require the least maintenance with maybe only even a one time treatment, but this is the high-end method and there will be probably others before Like any technology, when it first starts off, it will be a bit shaky, a bit risky, it will be very laborious and expensive, but there will be enormous market pressures that will result in progressive refinement and improvement to the technology so that it becomes more effective and convenient. HOW LONG COULD WE LIVE WITH THIS TECHNOLOGICAL ADVANCEMENT? Potentially forever, however, if you had never seen any mortality statistics, you might subscribe to what’s called the “lightning bolt theory” of mortality. In this view, death is the result of a sudden and unexpected event over which you have no control. It’s sort of an ancient Greek perspective: there are angry gods carousing carelessly overhead, and every so often they hurl a lightning bolt towards Earth, which kills you if you happen to be in the wrong place at the wrong time. These are the “lightning bolts” of disease and cancer and car accidents, things that you can escape for a long time if you’re lucky but will eventually catch up to you. Your probability of dying during a given year would be constant, and wouldn’t increase from one year to the next. From current statistics this probability for the populations of the UK and USA is somewhere around 0,05% per year (varying on location and lifestyle) Taking it to numbers with such mortality rate, out of 1 million people today, 950'000 would still be alive after 100 years, 780'000 after 500 years and around 600'000 in 1000 years. And then after ten thousand years — if you are lucky — you will be one of the 7000 people still alive. This is also a great argument for people who falsely fear that immortality will lead to overpopulation. As if all 7,6 billion people who life today would become immortal, after 18'000 years only 1 million of them would be still alive. This article was presented to you by Immortality Foundation.

Stories, News & Mentions about METANOVAS

Metanovas’s Stories on HackerNoon