Motivation
Nowadays, artificial intelligence (AI) is gaining increasing attention across a variety of research fields. Beyond using it in the computer game industry, for example to model non-player characters having their own behaviour and taking their own decisions, AI has shown tremendous impact in various applications found in healthcare, automotive, economics and many more areas.
What makes AI special
While conventional algorithms are very problem-specific by design, AI approaches are able to self-adjust to the problem without the need of explicitly modelled rules or assumptions. Such AI approaches are based on e.g. trained neural-networks, which can be used as a black-box for the users, lowering the amount of domain-knowledge required. On the downside, AI driven algorithms naturally involve a large number of computations, putting high demands on computation resources, such as dedicated AI rigs or clusters.
Hadron’s approach: many-as-one
Hadron promises to overcome this by setting up a network of AI-capable devices, spread across the whole world, to solve global AI problems in a distributed fashion. That means each device only needs to perform a small fraction of the AI calculations. Since Hadron builds on a javascript based neural network framework using GPU, every browser-enabled device supporting WebGL is eligible to contribute.
Computational power out of the pocket
With owning a smartphone, the majority of people have a powerful device in their pocket. Based on the assumption that smartphones are idle 90% of the time, they bear a large potential source of computational power to leverage for AI tasks.
Building on blockchain technology
By building on the blockchain, the network of AI devices becomes decentralized, avoiding a single-point-of-failure. Moreover, token economics encourage the contributors to continue allocating their computational resources for the AI tasks as they can be directly rewarded with tokens. With that, consensus-based decisions can be made, for example in form of a voting for the next global AI challenge to tackle.
Hadron’s current status and near-future
- Hadron is currently in the alpha phase with lots of early-registered miners already involved in the calculation of typical AI tasks.
- Each early miner gets rewarded with coins which will be of use in the future. Potential bugs are reported to the developers.
- Two remarkable partnerships have already been made: Bitlearn Network and AIKON.
- The beta phase will start in roughly a few month, registration is open via Bitcointalk Forum post (link below).
- An ERC-20 tokenization approach is considered after release.
- Development of an ecosystem around AI computation on the blockchain can be possible.
Updates
June
- HADRON is backed up by Neo Global Capital — the investment arm of the NEO Foundation. http://ngc.fund/
July
- HADRON collects donations from alpha miners and added personal funds on top to help the West Japan Disaster Relief with a total of 120 ETH.
- HADRON collaborates with Hubble’s Space Telescope Institute by classifying deep space images with artificial intelligence.
- HADRON is the first US blockchain project backed up by China’s top venture capital firm Matrix Partners.
More information and sign-up for the upcoming beta
Web: https://hadron.cloud/
BCT: https://bitcointalk.org/index.php?topic=2142232.0
Telegram: https://t.me/joinchat/GahFuELlfJoWE1vsFQf3VQ
Twitter: https://twitter.com/hadron_cloud
Facebook: https://www.facebook.com/HadronCloud/
Reddit: https://www.reddit.com/r/HadronCoin/
Team: https://hadron.cloud/#team-section
Whitepaper: https://hadron.cloud/whitepaper.html