Notice of the General Office of Beijing Municipal People’s Government on Printing and Distributing Several Measures for Promoting the Innovation and Development of General Artificial Intelligence in B
Beijing Zhengban made [2023] No.15
District People’s governments, municipal government commissions, offices and bureaus, and municipal institutions:
With the consent of the municipal government, we hereby print and distribute "Several Measures for Promoting the Innovation and Development of General Artificial Intelligence in Beijing" to you, please conscientiously implement it according to the actual situation.
General Office of Beijing Municipal People’s Government
May 23, 2023
(This piece is publicly released)
Some measures to promote the innovation and development of general artificial intelligence in Beijing
In order to implement the Implementation Plan for Accelerating the Construction of Artificial Intelligence Innovation Sources with Global Influence in Beijing (2023-2025), give full play to the guiding role of the government and the catalytic role of the innovation platform, integrate innovative resources, strengthen the allocation of elements, create an innovative ecology, pay attention to risk prevention, and promote the innovation-oriented and rational and healthy development of general artificial intelligence in this city, the following measures are formulated.
First, improve the overall supply capacity of computing resources
(A) the organization of commercial computing power to meet urgent needs
Give full play to the advantages of computing power resources in this city, implement the computing power partner program, speed up the collection of existing computing power through strengthening cooperation with cloud vendors, clarify the technical standards, software and hardware service requirements, computing power supply scale and support measures, provide diversified and high-quality inclusive computing power for innovative subjects, and ensure the computing power demand of artificial intelligence technology innovation and product development.
(2) Effectively promote the construction of new computing infrastructure.
Incorporate new computing power construction projects into the computing power partner program, accelerate the construction of Beijing artificial intelligence public computing power center and Beijing digital economy computing power center in Haidian District and Chaoyang District, form large-scale advanced computing power supply capacity, and support the research and development of large-scale language models, large-scale visual models, multi-modal large models, scientific computing large models, large-scale fine neural network simulation models, and brain-inspired neural networks with hundreds of billions of parameters.
(3) Building a unified cloudy computing power dispatching platform.
Aiming at the demand of flexible computing power, a cloudy computing power scheduling platform is built to realize unified management and unified operation of heterogeneous computing environments, which is convenient for innovative subjects to run all kinds of artificial intelligence computing tasks seamlessly, economically and efficiently in different cloud environments. Further optimize the basic optical transmission network directly connected between this city and Tianjin, Hebei Province, Shanxi Province, Inner Mongolia Autonomous Region and other regional computing power clusters, and improve the integrated dispatching capacity of computing power around Beijing.
Second, improve the supply capacity of high-quality data elements
(D) Collecting high-quality basic training data sets
Organize relevant institutions to integrate and clean Chinese pre-training data to form an open basic training data set with safety and compliance; Continue to expand multi-modal data sources, build high-quality pre-training corpora of large models such as words, pictures, audio and video, and support data circulation and trading in legally established data trading institutions.
(five) planning to build a data training base
Accelerate the construction of a pilot demonstration zone for the data basic system, explore the creation of a data training base, promote the high-level opening of data elements, and improve the scale and quality of the artificial intelligence data labeling library in this city. Internet platforms that provide content information services are encouraged to provide high-quality corpus data for innovative subjects to apply for. Explore commercial scenario cooperation based on data contribution and model application.
(6) Build a crowdsourcing service platform for fine labeling of data sets.
By crowdsourcing service, we will build a platform for fine annotation of data sets, develop an intelligent cloud service system, and integrate related tool applications. Encourage and organize professionals from different disciplines to participate in labeling multimodal training data and instruction data to improve the quality of data sets. Study the incentive mechanism of the platform to promote the sustained and benign development of the platform.
Third, the system builds a general artificial intelligence technology system such as a large model.
(seven) to carry out large-scale model innovation algorithm and key technology research.
Focusing on model construction, training, optimization and alignment, reasoning and deployment, we will actively explore the innovation of basic model architecture, study large-scale model efficient parallel training technology and optimization methods such as cognitive reasoning, instruction learning and human intention alignment, and develop efficient compression and end-to-end deployment technologies that support tens of billions of parameter model reasoning, form a complete and efficient technical system, and encourage the ecological construction of open source technology.
(eight) to strengthen the large model training data collection and management tools research and development.
Focusing on the training data "collection, storage, management, research and use", this paper studies the real-time updating technology of Internet data, the integration and classification method of multi-source heterogeneous data, builds the relevant system of data management platform, and develops algorithms and tools for data cleaning, labeling, classification, annotation and content review.
(9) Building an open service platform for large-scale model evaluation.
Encourage third-party non-profit organizations to build multi-modal and multi-dimensional basic model evaluation benchmarks and evaluation methods; Study the model evaluation algorithm aided by artificial intelligence, and develop a multi-dimensional basic model evaluation tool set including universality, efficiency, intelligence and robustness; Build an open service platform for large model evaluation, establish a fair and efficient adaptive evaluation system, and realize automatic adaptive evaluation of large models according to different goals and tasks.
(10) Building the basic software and hardware system of large model.
Support the development of large model distributed training system to realize efficient and automatic parallel training tasks. A new generation of artificial intelligence compiler suitable for model training scene is developed to realize automatic generation and optimization of operators. Promote the wide adaptation of artificial intelligence training reasoning chip and frame model, develop artificial intelligence chip evaluation system, and realize automatic evaluation of basic software and hardware.
(eleven) to explore a new path of general artificial intelligence.
We will develop the basic theoretical system for general artificial intelligence, strengthen the basic theoretical research on mathematical mechanism, autonomous collaboration and decision-making of artificial intelligence, and explore new paths of general artificial intelligence such as general agent, embodied intelligence and brain-like intelligence. Support the research of value-driven and causal-driven general agents, build a unified theoretical framework system, rating standards and test platforms, develop operating systems and programming languages, and promote the application of the underlying technical architecture of general agents. Promote the research and application of intelligent system with body, and break through the perception, cognition and decision-making technology of robots under complex conditions such as open environment, generalized scenes and continuous tasks. Support the exploration of brain-like intelligence, study the core technologies such as connection mode, coding mechanism and information processing of brain neurons, and inspire new artificial neural network model modeling and training methods.
Fourth, promote the application of general artificial intelligence technology innovation scenarios
(twelve) to promote the demonstration application in the field of government services.
Focusing on the work of government consultation, policy service, handling complaints immediately, and handling government affairs, we will make use of the advantages of artificial intelligence in semantic understanding, autonomous learning and intelligent reasoning to improve the intelligent question-and-answer level of the government consultation system, enhance the standardized management and accurate service capabilities of the "Beijing Policy" platform, assist the citizen service hotline to respond to citizens’ demands efficiently, and promote accurate guidance and efficient approval of government affairs.
(thirteen) to explore the demonstration application in the medical field.
Support qualified research-oriented medical institutions to refine the requirements of intelligent guidance, auxiliary diagnosis, intelligent treatment and other scenarios, fully tap medical literature, medical knowledge maps, medical images, biological indicators and other multimodal medical data, and develop intelligent applications with artificial intelligence innovation subjects to realize accurate identification and prediction of symptoms, signs and special diseases, and improve the intelligent level of disease diagnosis, treatment, prevention and whole course management.
(fourteen) to explore the demonstration application in the field of scientific research.
Develop scientific intelligence and accelerate artificial intelligence technology to empower scientific research in the fields of new materials and innovative drugs. Support related laboratories in the fields of energy, materials and biology to set up scientific research cooperation projects, carry out joint research and development with innovative subjects of artificial intelligence, fully tap experimental data in the fields of materials, protein and molecular drugs, develop scientific calculation models, carry out prediction of chemical structure sequences of new alloy materials, protein sequences and innovative drugs, and shorten the period of scientific research experiments.
(fifteen) to promote the demonstration application in the financial field.
System layout "unveiling the list" project to promote financial institutions to further open up industry application scenarios; Support financial science and technology innovation subjects to focus on intelligent risk control, intelligent investment, intelligent customer service and other links, research and develop financial professional long-text accurate analytical modeling technology, complex decision logic and model information processing integration technology, and support investment-assisted decision-making in the financial field.
(sixteen) explore the demonstration application in the field of automatic driving.
Support autonomous driving innovative subjects to develop multi-modal fusion sensing technology, based on vehicle-road collaborative data and multi-sensor fusion data of vehicle driving, improve the multi-dimensional perception and prediction performance of autonomous driving model, effectively solve the long tail problem of complex scenes, and help improve the generalization ability of vehicle-mounted autonomous driving model. Support the opening of vehicle-road collaborative autonomous driving data set in the construction of Beijing high-level autonomous driving demonstration zone 3.0 project. Carry out the cloud-controlled autopilot model test based on low-delay communication, and explore the new technical path of autopilot.
(seventeen) to promote the demonstration and application in the field of urban governance.
Support artificial intelligence innovation subjects to combine the needs of smart city construction scenarios, take the lead in applying large model technology in urban brain construction, accelerate the research and development of multi-dimensional perception system fusion processing technology, realize unified perception, correlation analysis and situation prediction of the underlying business of smart cities, and provide more comprehensive support for urban governance decisions.
V. Explore and create an inclusive and prudent regulatory environment.
(eighteen) continue to promote the innovation of regulatory policies and regulatory processes.
Explore and create a stable and inclusive regulatory environment, encourage innovative subjects to adopt safe and reliable software, tools, computing and data resources, and carry out independent innovation, popularization and application, and international cooperation in basic technologies such as artificial intelligence algorithms and frameworks. Strive to try first in the core area of Zhongguancun National Independent Innovation Demonstration Zone and promote the implementation of inclusive and prudent supervision pilot.
(nineteen) to establish a normalized service and guidance mechanism.
For Internet information services related to artificial intelligence with public opinion attributes or social mobilization ability, we will carry out normalized contact services, and guide innovative subjects to introduce technical tools for safety testing, declare safety assessment according to regulations, and perform procedures such as algorithm filing.
(twenty) to strengthen the network service security protection and personal data protection.
Guide innovative subjects to strengthen network and data security management, implement the main responsibility of network security, data security and personal information protection, and strengthen the construction of security management system and the implementation of work. Encourage innovative entities to carry out data security management certification and personal information protection certification, implement data cross-border transmission security management system, and comprehensively improve network security and data security protection capabilities.
(twenty-one) continue to strengthen the ethical governance of science and technology.
Strengthen the research on artificial intelligence ethical safety norms and social governance practice. We will build a public service platform for ethical governance of science and technology in the field of general artificial intelligence to serve government supervision and promote industry self-discipline. Conduct scientific and technological ethics review and related business training, and strengthen the awareness of scientific and technological ethics norms of all responsible subjects. Carry out in-depth education and publicity on science and technology ethics, and build a good atmosphere of artificial intelligence science and technology ethics.