2012年1月Chinese voice assistant intelligence 360 was released.
2013年1月Released the semantic open platform of artificial intelligence-semantic cloud, which is oriented to the field of APP.
2013年6月Smart 360 has exceeded 10 million users and become a third-party voice assistant with over 10 million users.
2014年6月The artificial intelligence speaker-Xiao Zhi Super Speaker (released by Amazon Echo in December 2014) was released, equipped with a semantic cloud brain.
2014年7月Become a member of the board of directors of China Electronic Audio Association.
2015年12月Xiao Zhi housekeeper robot was released, equipped with Semantic Cloud Brain 2.0.
2015年12月Xiao Zhi housekeeper robot was released, equipped with Semantic Cloud Brain 2.0.
2016年3月Xiao Zhi housekeeper robot and HTC Vive virtual reality helmet won CHINABANG award 2016 Science and Technology Innovation Award.
2017年4月Brand-new upgrade semantic cloud, released DeepBrian brain platform, brand-new upgrade.
2017年5月Start the blockchain research and project of cloud brain artificial intelligence operating system; Establish a blockchain laboratory and develop based on blockchain application architecture.
2017年5月Start the blockchain research and project of cloud brain artificial intelligence operating system; Establish a blockchain laboratory and develop based on blockchain application architecture.
2017年5月Start the blockchain research and project of cloud brain artificial intelligence operating system; Establish a blockchain laboratory and develop based on blockchain application architecture.
2017年6月Received angel investment totaling RMB 32 million from Jinshajiang Venture Capital, Gobi Investment and Qianshi Investment.
2017年7月1. The HTML5 version of 1.DeepBrain brain platform skill store is released, and consumers are free to add skills to AI hardware; 2. Version 0.6 of the White Paper on Deep Brain Chain Token Sales was released.
2017年8月Won the first place in the evaluation of SMP2017 Chinese man-machine dialogue by Harbin Institute of Technology AI; More than 20 domestic top artificial intelligence teams participated in the competition; Version 1.0 of the White Paper on Deep Brain Chain was released.
2017年8月Won the first place in the evaluation of SMP2017 Chinese man-machine dialogue by Harbin Institute of Technology AI; More than 20 domestic top artificial intelligence teams participated in the competition; Version 1.0 of the White Paper on Deep Brain Chain was released.
2017年9月1. The R&D team moved to the new office of Hongqiao Tianjie in Shanghai Dahongqiao Business Circle; 2. Invited to participate in the China AI 30 closed-door forum held in Beijing and the internal seminar on interpreting the State Councils "Notice on the Development Plan of a New Generation of Artificial Intelligence"; 3. Invited to the Digital Assets Summit.
2017年第四季度In January and October, he was invited to participate in the "First Global Financial Technology and Blockchain China Summit 2017"; 2. The first phase of the deep brain chain based on the NEO contract is completed and tokens are issued; 3. Support the withdrawal of money, and DBC assets are launched on third-party exchanges; 4. The quarterly key information disclosure of the project; 5. Deep brain chain promotion and vendor access.
2018年第一季度Complete the development of the underlying framework and core components, complete the development of DBC AI test network, and build an internal test environment; Support internal AI test users to submit AI training requirements to DBC AI test network, and complete training tasks through DBC AI test network; Development of DeepBrain Chain crowdsourcing system for community tasks in official website: End the first round of global meetup (Dublin, Hamburg, Amsterdam, San Francisco).
2018年第二季度Complete the development of AI user management, group management, role management and authority management; Complete the integration and development of DBC AI test network and blockchain system; Complete the construction and testing of DBC internal network environment; On June 30, Skynet tested the online line; Establish a silicon valley research and development center.
2018年第三季度Further improve and test the blockchain network and start the network internal test of key users; Support the integration and deployment of multiple deep learning engines; Support model algorithm publishing and selecting published model algorithms, and publish training/testing tasks after packaging data; Support the monitoring and statistical analysis of DBC network; Support AI users to perform AI calculation in DBC network; Support DBC network security requirements; On August 8, Skynet officially launched online, and DBC can be used to purchase GPU computing power.
2018年第四季度Further improve and test the blockchain network and start the network public beta; Further support the integration and deployment of multiple deep learning engines; Support AI users scoring of task execution; Support API to open to third-party users; Further improve DBC network security.
2019年第一季度Further improve and test the blockchain network and make modifications according to user feedback and community needs; Support AI data trading and evaluation; Support AI data encryption and privacy protection.
2019年第二季度Complete GPU isolation function; Design and develop new dbc message encryption function and reconstruct dbc message signature function; Use ECDH algorithm to generate a one-time share secret, and then use share secret to encrypt the log message with CBC-AES. Integrate dbc message encryption and dbc message signature. Design, develop and complete offline certification of training tasks to solve the problem that mining opportunities are used for free when auth WEB fails; Design and optimize the dbc matrix protocol, and increase the enable_fast_forward option.
2019年第三季度Transparent forwarding of optional parameters is supported, and the current dbc intermediate node will discard the unrecognized fields when forwarding messages; Support transparent forwarding of unknown types of messages; Reconstruct dbc source code and prepare open source documents; Complement the internal design documents of restful API.
2019年第四季度Provide a standard template for creating GPU Server, which corresponds to the following scenarios: (a) GPU Server uses the public IP provided by Computing Node; (b) GPU Server uses ngrok proxy；; Optimize the creation of GPU Server: reduce the time-consuming of GPU Server creation from several minutes to tens of seconds by the following means: (a) install the startup code of the standard template mentioned in 2.2 to the computing node and copy it to each GPU server; (b) Using the standard template will skip the installation, startup and code download of IPFS; The functions of memory management, CPU management and hard disk management of the container are developed. Completed the GPU adding and unloading functions of the container.
2020年第一季度The functions of GPU container and CPU container are officially launched; The automatic verification function of adding machines is online. After the miners add machines, the system will automatically verify the availability of the machines. The verification can be automatically online, and the whole process from the deep brain chain payment network to the AI computing network to the cloud platform can be fully verified without any human intervention. The container image supporting TensorFlow2.0 is online, and users can use it directly without installing TensorFlow2.0 themselves. The function of private network disk is officially launched, which makes it convenient for most small white users to upload files to GPU server; The mirror image supporting drug molecular dynamics simulation is launched, which is convenient for the majority of drug research and development workers to use GPU computing power.
2020年第二季度Add tensorboard support to facilitate AI users to visualize during training; SSH verification mechanism of mining machine online is added; The cloud platform server has been secured to prevent data loss; On-line GPU machine classification mechanism, according to memory space and hard disk space to classify machines; Improve the display of machine rental notice information, which is convenient for miners to check the machine income conveniently; A payment verification mechanism is added to prevent payment failure caused by network problems; After the development, the mirror image is automatically obtained, and a new version of the mirror image is made, which adds restrictions on the absolute value of CPU cores in CPU containers to prevent malicious mining from wasting its own resources; Improve the online and offline mechanism of the machine; Added hard disk full restart limit and hard disk space shortage prompt mechanism; The switching mechanism between GPU container and CPU container is improved, which no longer depends on the data size in the container, greatly improving the experience. The original 100g data takes 1 hour to start the container, and the new version only takes 3 minutes to complete; The distributed computing network layer adds a script startup mechanism; A new version of the image is made, the python execution environment is unified, and the files directory of jupyter and private network disk is unified.
2020年第三季度An automatic recovery mechanism is added when the switching between GPU container and CPU container fails. The automatic recovery mechanism of disconnection of jupyter and private network disk ngrok is added; The automatic recovery mechanism of abnormal termination of ssh, jupyter and private network disk processes is added; Developed and produced brand-new mirror images of tersorflow2.1, tersorflow2.0, tensorflow1.14, pytorch1.2, pytorch1.4, pytorch1.5, mxnet1.5 and mxnet1.6; The deep learning image has been completely upgraded, and a lot of junk content has been deleted, making the image more lightweight; Added mxnet1.5, mxnet1.6, tensorflow2.1, pytorch1.5 and the latest molecular dynamics environment support; The corresponding version of the ipykernel kernel is added to each image, which is convenient for jupyter to use. Automatically activate the virtual environment of the corresponding version of the mirror image, which is more convenient to use; Jupyter of unified virtual environment and python version under terminal access; Mirror Ubuntu version to version 18.04; The compiled version of opencv is pre-installed in the image, which greatly improves the experience of opencv. Added docker automatic detection and restart function; Added detailed server usage documents and help documents; Optimize the response mechanism of AI computing network; The deep learning mirror and DBC network are upgraded, and the compiled version of opencv is pre-installed in the mirror, which greatly improves the experience of opencv. Added the default download source of pip as Tsinghua source; Optimize the response mechanism of AI computing network; By default, the support of cuda9 is added; The calculation mirror of FileCoinSnake is added to support API interface access for c1 and c2 calculation of filecoin seal. Added support for Filecoin snark GPU computing.
2020年第四季度Added the discount function of monthly, quarterly and annual subscription; Added the function of supporting whole rent; The AI reasoning area has been added, and the machines in this area can be rented, and additional open ports can be used; The website interface layout has been revised again, and the classification is clearer; The DBC super node leasing function is added. If you want to participate in the DBC super node, you can rent the machines in the DBC network by pledging a certain number of DBCs without buying the machines yourself. The pledge function is added. If you want to enter the highly stable AI training area, you need the machine to run stably for at least 144 hours and pledge DBC；. Increase that functions of the upper rack machine and the low rack machine; Optimized GPU acceleration algorithm; Optimize the cloud platform wallet and increase the gas function; Study the combination of privacy computing and DBC network; Optimize the dynamic expansion function of DBC network nodes; Upgraded the deep learning mirror that supports deepshare courses and paper; In order to cope with the continuous increase of machines in DBC network, the dynamic expansion function of DBC nodes is further optimized; Go online with AIM machines. Now all machines are online and can be delivered to AIM buyers at any time. Further optimize the message forwarding function of DBC network nodes; According to the needs of cloud games and rendering users, a scheduling engine based on GPU virtual machine is developed. Modifying web wallet based on substrate; Support nearly 30 GPU cloud platforms based on deep brain chain.
2021年第一季度As soon as the public beta phase of DBC main network is started; Developing an online reward contract for computing nodes; Develop a lease and destruction contract for computing nodes.
2021年第二季度The second phase of DBC main network public beta is started; The first stage of online reward contract testing for computing nodes; The main network of DBC is officially launched, and the token exchange DBC is started; The online reward contract for computing nodes was officially launched.
2021年第三季度The lease and destruction contract of computing nodes was officially launched, and DBC began to enter the era of deflation.
2021年第四季度Upgrade and improve smart contracts; Support AI training and reasoning, cloud games, graphic rendering, ETH2.0 and other multi-scenario customer advanced customization needs.
