Decentralized AI computing enterprise, DeepBrain Chain (DBC), officially announced the activation of their Skynet Project, recruiting AI computing power from all over the world, building a globally decentralized AI TestNet to strengthen the foundation for the MainNet.
DeepBrain Chain CEO Yong He said:
“The activation of Skynet Project means that DeepBrain Chain’s AI TestNet has entered large scale beta testing. Building on blockchain and decentralized AI computing, DeepBrain Chain will help AI enterprises take off globally, lowering their computing cost, accelerating their products’ time to market.”
The “Skynet Project”
The word ”Skynet” usually references the conscious AI super computer that plans to destroy human being in the sci-fi film “Terminator.” DeepBrain Chain’s Skynet Project has steered the interest of the global AI community.
On June 3rd, DeepBrain Chain’s R&D team has successfully ran 3 real-life AI training models on their TestNet for the first time. In order to assess TestNet’s function comprehensively, DeepBrain Chain has activated its Skynet Project, recruiting AI computing power from all over the world.
Computing power, algorithms and models are the three crucial elements of AI’s development, one cannot go without the other. Computing power cost has been a huge pain point for AI enterprises, and this is where DeepBrain Chain excels. There are many platforms that offer AI cloud computing, but they are all costly. For small to medium size AI enterprises, on average 20%-30% of their funds have to be used on buying AI computing power, DeepBrain Chain can help them reduce that cost by up to 70%.
For university research centers and AI enterprises, they are AI computing consumers but also providers. Because different enterprises need different amount of computing power, they also use the computing power at different times, this leads to some of that computing power being idle at times.
Benefits include having the priority to join DeepBrain Chain’s MainNet as computing nodes; higher chances at being elected as consensus nodes; and use the computing resources from other TestNet nodes on the platform for free. This is a one-time chance with a small investment.
Multiple Use Cases Of DeepBrain Chain’s Platform
At the moment AI applications have spread into every corner of our daily lives. Driverless cars, face recognition and Natural Language Processing are all part of this AI invasion.
Different to common cryptocurrency mining machines, machines that provide AI computing power has extremely high configurations, as well as being highly customized. Usual CPUs and GPUs in personal computers cannot compare, nor do they have the software to run AI computing. The AI computing power DeepBrain Chain is recruiting in this campaign has to pass a certain threshold. The machines have to have GPUs with NVIDIA’s 1080Ti GPU or above. DeepBrain Chain plans to select 1,000 AI computing machines to be test nodes for the TestNet.
DeepBrain Chain is building a decentralized AI public chain, offering any enterprises or individual with AI needs high-performance computing power and data privacy. AI enterprises can also issue their own token on DeepBrain Chain, deploy their products on the platform. Every AI computing node on DeepBrain Chain earns income in two ways, the DBC token that the system rewards them for mining and the DBC token paid for by AI computing power consumers. On the foundation of AI public chain, using the unified AI cloud network to build an AI model development and training platform as well as marketplace is the goal.
DeepBrain Chain’s Senior Test Architect & Specialist, Steve Miao said, “compared to our competitors, DeepBrain Chain offers powerful computing power while maintaining a low cost. Enterprises can run their common or optimized AI models, AI data on the platform. DeepBrain Chain’s network is faster at creating new blocks and has a shorter reaction time compared to BTC and ETH, offering users different levels of customized security solutions as well. Our comprehensive AI ecosystem and solutions are not seen in other suppliers.”
DeepBrain Chain can now support TensorFlow DeepLearning, H2P Machine Learning framework on their TestNet. TensorFlow is the industry’s most popular AI DeepLearning framework developed by Google, it can be used on image recognition, voice recognition, NLP and other various deep learning fields; H2O is the industry’s most popular machine learning and predictive analytics framework. These frameworks have the most AI and machine learning developers in the industry.
Protecting AI Enterprises’ Data
Data security is the lifeline of enterprises, DeepBrain Chain cares about the system’s security. For real-life AI training tasks, the AI training task submitter can track the status of the training at all times. Time of computing and diaries will be saved, if a computing node behaves maliciously their deposit will be taken away, and the system will bring in system ranking and user review mechanism, conduct ranking on the result of training.
Also, DeepBrain Chain’s AI training tasks have a categorizing system, normal level tasks will get computing nodes assigned by the system randomly (to prevent malicious behavior of manually accumulating higher ratings by continuing to run low level training task to get good ranking). For high level tasks, the task submitter can choose nodes they already trust to do part of the training, as well as verifying the result of the training.
Steve Miao said, “in the future, whoever has more powerful computing power will be the terminator in the film, DeepBrain Chain consolidates all computing power to service the industry. Our service perfectly serves NLP, image recognition, voice recognition and other applications in the AI industry.”
According to his explanation, in the future DeepBrain Chain will become a “docker” with unlimited possibilities, you will be able to run other mining programs on the platform. It can be a P2P type Web service node or something else, the possibilities are endless.
DeepBrain Chain’s TestNet plans to go live at the end of June. After their MainNet goes live in the second half of the year, the R&D team will continue to research the possibility of large scale AI distributed training and the whole network wise distributed training.
At the Global Blockchain Leaders Summit (GBLS) held in Hangzhou on June 6, 2018, DeepBrain Chain was awarded “Most Outstanding Technology Award” because of their innovative business model using blockchain technology, solving the three biggest hurdles, high computing cost, low data security and long product-launching period in the AI industry.
Source: PR Newswire