Dear DeepBrain Chain Community,
Welcome to the latest fortnightly progress report. We recently saw the 10th anniversary of the bitcoin whitepaper, which introduced the first cryptocurrency to the world. This anniversary reminds us of the potential for innovation to spur revolutionary change, and fills us with ambition and inspiration for our own vision at DeepBrain Chain. The past two weeks have seen a primary focus on running AI tasks on our Training Net, with our first clients providing feedback with which we can devise upgrades and new iterations to maximize performance. We will be detailing some of this feedback in this report. DeepBrain Chain announced a formal cooperation agreement with Anhui Sanlian University, a renowned educational institute in China with over 18000 students. Our Chief Technical Officer, Bruce Feng, will be presenting our latest technical developments, including an upcoming major upgrade to the network, in a DECO livestream on the 15th of November. We also announced that we will be attending the Bitfish AI Night in Bangkok, Thailand on the 28th of November with Cortex.AI, and we will be discussing the progress of the platform in detail with our strong local community, as well as taking the opportunity to meet with industry experts and potential clients.
AI Training Net Tasks, Feedback and Upgrades
The recent weeks have seen a strong focus on client and user feedback for the AI Training Net. We currently have over 250 tasks that have been run on the network and have received a lot of user feedback that we can implement moving forward. Once the first round of AIMs join the network before the end of the year it will be vastly expanded and we expect both serious computational power and serious demand for the network. Later this month, we will be rolling out a major overhaul of the network that will fix a number of issues and generally upgrade all-round performance. CTO and Chief Architect Bruce Feng had this to say:
‘’We have optimized our Skynet network according to existing AI users feedback. We will be supporting mining machines to allow them to conduct other token mining tasks when not running AI Training to maximize the utilization of mining machine resources on the network. We also aim to optimize the AI user experience on our official website. We will support remote log-in capabilities for AI users to make adjustments to their mining machines online, optimizing their developing experience. ‘’
Some examples of user feedback:
(Disclaimer: some clients and users wish to remain anonymous and as such we have excluded their identity in the interest of protecting their privacy)
- Anonymous Chinese technology company from Beijing
‘’My user experience of this system is great in general! In regards to the user manuals, if you could include more details in order to offer a more complete guide it would be even better. You might want to consider making different manuals for Windows, Mac and Linux so it’s easier for users to read if they run different operating systems. ‘’
We are currently working on user guides that will include a number of different operating systems to allow for nuanced instructions that guide users through their respective steps.
2. Anonymous user that used Tensorflow, Pytorch, Caffe and Mxnet frameworks to train AI models
‘’For payment, AI users are confused when things involve cryptocurrency (e.g registration, linking your wallet, buying DBC etc). They may be rather unwilling to try these things so you need to devise easier methods and have them clearly explained’’
Our short term solution is to offer technical support services that will complete these steps for them as requested. Long term, we are looking at the inclusion of a fiat payment system and the use of a third party service to purchase DBC from the market with that fiat. Solutions we are looking to integrate through the official DBC website include automating the linking of the ID of a client’s machine that they are using for training with their registered account. We will also be updating the steps for selecting your mining machine and GPU information on our official website to allow easier access to your desired training specifications.
- Regarding uploading documents through IPFS;
A) Can I delete a document after it is uploaded?
At the moment we do not have a manual delete function, though one may be implemented in the future. For now, our system will automatically delete data that has not been used for a long period of time.
B) Can I download a file that I have uploaded?
Yes, use the following commend line:
ipfs get QmTtTjgbA88MKDNktipTN4aVF79NRtUqgsAAAGTV3wFqZU o pilot_dataset
C) Once a compressed file is successfully uploaded to your server, how can I unzip it?
Please follow instructions here.
- Regarding GPU specification
Can I specify which GPU to use in my script and run two tasks at the same time?
Yes, just specify as such in your program.
We are constantly updating our Github and Github Wiki with solutions and changes as we upgrade the network. To have a look at more feedback and get an idea as to how the general experience of the AI Training Net has been so far, have a look at these Github links.
DeepBrain Chain signs strategic co-operation agreement with Anhui Sanlian University
On October 30, 2018, the DeepBrain Chain Foundation and Anhui Sanlian University formally signed a strategic cooperation agreement. The two parties will integrate their respective resources and expertise to jointly utilize the DeepBrain Chain AI platform to develop AI solutions and applications. Furthermore, the two parties will work together to create a vocational training platform in the field of artificial intelligence and blockchain services for local students. Dr. Haisong Gu, co-founder of the VisionX project, which was incubated by the DECO, accepted a letter of appointment and became a visiting professor at Anhui Sanlian University. We are very excited about this cooperation, as we know the education sector is a huge market that can greatly benefit from the services our platform is able to provide. We know that this is only the beginning of further institutional cooperation to come, and we look forward with earnest enthusiasm to a mutually beneficial relationship.
CTO and Chief Architect Bruce Feng to present from China
Chief Technical Officer Bruce Feng will be speaking via live broadcast on the 15th of November. The initial broadcast will not be in English but the information will be translated shortly after. Bruce will also be addressing our English-speaking community soon after the initial broadcast. As the lead architect for the DeepBrain Chain platform, Bruce hopes to communicate with investors in detail about the progression of the technology side of the project, advancements in which have been substantial. As we prepare for the eventual launch of the mainnet, communicating with our investors about the more technical aspects of DeepBrain Chain is extremely important to us, and this is why we have chosen this time to allow our Chief Technical Officer the opportunity to present to the global community. Expect more details very soon.
DeepToken Exchange Wins Top 10 Place in Houbi Cloud Trading Competition
The first round of exchange trading competitions with Huobi Cloud has successfully drawn to a close on October 30th. DeepToken Exchange, the World’s First AI Industry Digital Asset Exchange, has taken a place in the Top 10 for this competition. The trading competition was available to all exchanges operating on the Huobi Cloud system. Huobi boasts a massive global network, collaborating with over 100 companies and strategic partners internationally. The competition ran for a duration of 20 days, commencing on the 10th of October and ending on the 30th of October.
With this great success among fierce international competition, DeepToken Exchange has proven its ability to generate large amounts of trading volume even in its early stages. DeepToken Exchange will continue to exemplify optimal exchange operation with its integration of innovative operational practices such as communal voting, zero listing fees and the support of a global network of AI experts scrutinizing the rapidly expanding industry for only the highest quality projects to the great benefit of investors. Such scrutinization is particularly pertinent with the rate of development for an emergent technology like artificial intelligence, and DeepToken Exchange employs the support of partner DeepBrain Chain’s already robust community of experts and enthusiasts to ensure that every project that secures listing on the exchange is held to the highest standards of integrity and potential.
Beyond Blocks Blockchain Week AI Night
On the 28th of November, COO Maya will be attending the Beyond Blocks Blockchain Week AI Night with Bitfish and Cortex in Bangkok, Thailand on the 28th of November. Topic of the event is the integration of AI+Blockchain. Presenting with other AI projects gives us a chance to explore the different approaches we are developing to bring about the future of AI and the technological revolutions it will allow across all industries. We will also enjoy the chance to meet with our local community, whose support has been phenomenal.
DeepBrain Chain is now live on Delta Direct. All the latest updates will now be directly available on the Delta Direct app. Make sure you have notifications enabled so you never miss out.
Key tasks of the Shanghai R&D team:
DBC Skynet network optimization; v0.4 main chain refinement design;
- Network architecture design of multi-chain + fragmentation architecture: v0.4 main chain overall architecture refinement scheme design
- DGP solution analysis: read and analyze the DGP source code implemented by related technologies
- Data security solution analysis: container enhancement technology
- Data security solution analysis: decentralized encrypted tunneling scheme
- Blockchain network security analysis and design
- Official website iteration 3 testing
DBC Skynet Network Optimization:
- Compress communication function for reducing network bandwidth by 50% is being tested
- Support DBC network idle task scheduling
- Twitter binding problem
- Prompt update of node ID on official website
- Support miners (users) with viewing their mine revenue, income and expenditure details
- Support administrators to settle revenues
- Support miners to review mining machine audit opinions and resubmit mine audits
- Prompt information optimization for AI users
- Support DBC network health status detection
- Supports starting AI training tasks through command line parameters
- Support remote login container for online debugging
- Optimization of purchase interface for AI users
- Support users with setting BTM, RVN, DCR, XVG, ZEC addresses
- Support email notification results for the mining machine audit
- Support new user trial accounts for AI Training Net on DBC website
AI Training Net Data
- Number of GPUs: 70
- Machines used in training as requested by clients on the network: 183
- The cumulative number of AI training tasks performed: 288
(once the first round of AIMs are connected before the end of this year, the network will be vastly expanded. With more formidable computational capacity we expect and can cater for even greater demand on the platform)
Back-end Web Engineer
Aloys is very adept at web development, with over 6 years experience. Before joining DeepBrain Chain, Aloys worked for well-known financial companies including Tencent-backed CreditEase, Mobanker and more across multiple industries.
Global Technical Support
Bright will be responsible for providing technical support to our global community, putting his impressive proficiency in English, French and Arabic to good use. Bright boasts 12 years working experience in the industry; previous work at Huawei including Technical Development and Public Manager for a North African Data Center.
Front-end Web Developer
Kevin joins DeepBrain Chain as part of the expanding web team. He has 3 years of experience in front-end web development and 2 years of experience at renowned finance company, Mobanker.
We are hiring development engineers, security architects and other positions. All self-provided/recommended resumes are given feedback within 48 hours, and we welcome external referrals.
If you recommend a friend/classmate/colleague successfully, after the month of the candidate’s entry you will be given a recommendation reward of 20,000RMB.
Contact email: email@example.com
Please attach the resume of your recommendation, as well as referee contact information, so that we can issue bonuses as awarded.
DeepBrain Chain Global Ambassadors Program
Want a platform to give voice to your opinion? Want to be part of the decentralized team that steers DeepBrain Chain towards its future? Interested in displaying and developing your creativity and ingenuity to increase understanding and participation in the DeepBrain Chain ecosystem? Sign up to the Global Ambassador program using the link below!
Act quickly to be eligible for bountiful rewards!
Telegram DeepToken Exchange Group:(https://t.me/DeepTokenExchange)
Telegram English Group: (https://t.me/deepbrainchain)
Telegram Korean Group: (https://t.me/DeepBrainChainKor )
Telegram Vietnam Group:( https://t.me/DeepBrainChainVietnam)
Telegram Indonesia Group:(https://t.me/DeepBrainChainIndonesia )
Telegram Thai Group: ( https://t.me/DeepBrainChainThai )
Telegram Russian Group:(https://t.me/DeepBrainChainRussia )
Telegram Turkish Group: (https://t.me/DeepBrainChainTurkey)
Telegram Mining Machine Group:(https://t.me/DeepBrainChainAIminers )
About DeepBrain Chain
DeepBrain Chain is the world‘s first AI computing platform driven by blockchain. It uses blockchain technology to help AI companies save up to 70% of computing costs while protecting data privacy in AI training. Its vision is to build a “Decentralized AI Cloud Computing Platform” and become “The AWS in AI”
the DeepBrain Chain Team
Source: Deep Learning on Medium