CNCC2024
Brief introduction of the forum:
Young Scholars Development Forum: How to Face the Research Opportunities and Challenges in the Era of Large Models?
Time: 13:30-17:30, October 24
Location: Classroom 10, 1st Floor, Summer Garden-United Kingdom Pavilion
Note: If there is any change, please refer to the final information on the official website (https://ccf.org.cn/cncc2024).
At present, the explosive development of large models has had a profound and extensive impact on both industry and academia. In the industrial world, the large model provides strong technical support for the implementation of intelligent applications, promotes the comprehensive upgrade of automation and intelligence, and significantly improves production efficiency and innovation capabilities. In academia, large models not only broaden the research boundaries in the field of artificial intelligence, but also promote in-depth interdisciplinary collaboration, bringing unprecedented opportunities for academic research. At the same time, the rapid development of large models has also accelerated the integration of basic research and applied research, and promoted the all-round development of artificial intelligence technology from theory to practice.
However, the development of large models not only provides researchers with abundant research opportunities, but also poses new challenges. Especially young scholars, in the face of the opportunities and challenges brought by large models, how to maintain academic determination, how to carry out research innovatively, and how to make their own characteristics? It has become an important topic that young scholars need to think about urgently. In order to deeply discuss the opportunities and challenges faced by young scholars in the era of large models, this forum specially invited outstanding experts and scholars in various fields in the theory and application of large models to share their experiences. They will share real-world experiences on how to conduct innovative and impactful research with large models, and exchange insights, challenges, and insights along the way. The invited experts from multiple research fields shared the computing power integration problems of large models, the learning paradigm of knowledge exploration mechanisms and decision-making models, and the construction of large code models and machine language models.
This forum is initiated by the CCF 218 Club. The CCF 218 Club is mainly composed of scholars who have been selected for the "CCF Young Talent Development Program". The CCF Young Talent Development Program aims to support a group of young scholars who have achieved excellent results and have strong innovation capabilities, and the selected scholars are under the age of 32. This forum focuses on the development of young scholars, and aims to form useful suggestions for the development of young scholars through academic reports and exchanges.
Forum Agenda
| order | topic | Keynote speaker | unit |
| 1 | Software automation based on large models | Li Ge | Peking university |
| 2 | Large Machine Language Models: Technologies and Applications | Zhang Chao | Tsinghua University |
| 3 | Exploration of the knowledge mechanism of large models | Han Xianpei | Institute of Software, Chinese Academy of Sciences |
| 4 | Data-centric computing architecture: the integration of computing power in the era of large models | Wang Ying | Institute of Computing, Chinese Academy of Sciences |
| 5 | Reinforcement learning in the era of large models | Hao Jianye | Tianjin University |
| Panel link | All Forum Speakers |
Introduction of the chairman and guests of the forum
Chair of the Forum
civilization
Executive member of CCF 218Club, associate professor of Huazhong University of Science and Technology
Associate Professor, Master's Student/Doctoral Supervisor, School of Cyberspace Security, Huazhong University of Science and Technology, mainly focusing on software security, software testing and analysis, and code model security, etc., has published more than 50 CCF-A recommended conferences or journals in the field of software engineering, and won the ACM Rising Star Award (Wuhan), "Huazhong Scholar" Outstanding Young Scholar, Wuhan East Lake Scholar and other honors, and was selected into the 7th China Association for Science and Technology Young Talent Lifting Project Program.
Forum Speaker
Li Ge
Deputy Director of CCF Software Engineering Committee, Tenured Professor, School of Computer Science, Peking University, Changjiang Scholar of the Ministry of Education (Boya Distinguished Professor)
Member of the Standing Committee of CCF System Software Committee and Standing Committee of CCF Large Model Forum. Research interests: intelligent software development technology, key technology of intelligent software system, deep learning. He is a pioneer researcher of "deep learning-based program understanding and generation" in the world, and many of his achievements have been recognized as "pioneering achievements" by international scholars and widely cited. The research team led by him has maintained the international leading results in a number of research tasks, and is an internationally renowned research team in this field.
Title: Large Model-Based Software Automation
Report Summary: Large model technology has had a profound impact on software development technology. What is the software development assistance ability of the current large model? How will the future of software development change? The Program Understanding and Generation Research Team of Peking University is an early pioneer and continuous contribution team in the field of program understanding and generation based on deep learning. Based on his own research experience, the speaker gave a brief overview of the research process and development status of large-model-based program understanding and generation methods, and discussed the impact of large-model-based software development automation.
Zhang Chao
Executive Chairman of CCF 218Club, Vice Dean and Tenured Associate Professor of Tsinghua University Network Research Institute
He is a distinguished member of CCF, vice president of the Institute of Network Research of Tsinghua University, a tenured associate professor, a Huawei Named Professor, and a coach of the Blue Lotus team. He has won the honors of Tsinghua University Academic Newcomer and National Young Talent. His research interests include software and system security and artificial intelligence security. Developed the world's first large machine language model MLM.
Title: Large Machine Language Models: Technology and Applications
Report Summary: Software is facing security threats such as vulnerabilities, malicious code, and cracking, and a large number of software cannot obtain the source code, so binary security analysis capabilities are urgently needed. However, binary programs are difficult to analyze because they lack a lot of information compared to source programs. Large language model technology has brought a new breakthrough to binary program analysis. This report will share the Machine Language Large Model (MLM) solution, demonstrate the breakthrough capabilities and application effects of MLM in binary program security analysis, and discuss the development direction of software security with you.
Han Xianpei
Researcher at the Institute of Software, Chinese Academy of Sciences
He is a researcher at the Institute of Software of the Chinese Academy of Sciences and the deputy director of the Chinese Information Processing Laboratory, with research interests in natural language understanding, large models and knowledge graphs. He has undertaken more than 10 projects such as the strategic guidance of the Chinese Academy of Sciences, the 2030 project of scientific and technological innovation, and the national key research and development project. He has published more than 60 papers in important international conferences such as ACL, SIGIR, IJCAI, etc. He was selected as a member of the National Excellent Youth Program, the Young Talent Lifting Program of the China Association for Science and Technology and the Beijing Zhiyuan Young Scientist, and served as the director of the China Chinese Information Society and the deputy director of the Language and Knowledge Computing Professional Committee. The related achievements won the first prize of the Hanwang Youth Innovation Award and the first prize of the Science and Technology Award of the Chinese Chinese Information Processing Society.
Title: Exploration of the knowledge mechanism of large models
Abstract: In recent years, large models have shown strong capabilities, but the mechanism behind them is still unclear, which greatly restricts the improvement and application of large models. This report introduces some explorations of the research group in the knowledge mechanism of large models, such as the influence of knowledge on large models and the impact of large models on the external knowledge ecology, and discusses the relationship between large models and knowledge.
Wang Ying
Secretary-General of CCF Integrated Circuit Design Committee, researcher of Institute of Computing, Chinese Academy of Sciences
His main research interests include integrated circuit design automation, object artificial intelligence system, energy-efficient chip design and storage system design, and he has presided over the key research and development projects of the National Science and Technology Committee. He has published more than 100 CCF-A papers in the field of integrated circuits and system structures. He has won the CCF Young Scientist Award, the First Prize of CCF Technology Invention Award, the IEEE/ACM DAC Innovation Award under 40, and the IEEE TC Best Paper of the Year.
Title: Data-Centric Computing Architecture: Computing Integration in the Era of Large Models
In order to achieve this goal in the current slowdown of advanced industrial development, the expansion of computing power through the scale-up and scale-out route has to face the constraints of data communication walls and storage walls, so the data-centric computing architecture has become a new architecture with high hopes in the AI era. In order to alleviate the problem of power wall and storage wall, whether it is the integration of storage and computing based on new devices, SRAM and DRAM-based near-memory computing, or flash-based near-data computing systems have become a research hotspot in academia and industry, the report will introduce the advantages and combination of different storage and computing or near-data computing architecture solutions in differentiated application scenarios, and focus on how different storage and computing architectures can combine heterogeneous computing and advanced integration technology to cope with the explosive growth of system storage and interconnection resources in the era of large models.
Hao Jianye
He is an associate professor at Tianjin University and the director of Huawei's Noah's Ark Decision Reasoning Laboratory
Director of Huawei Noah's Decision Reasoning Lab and Associate Professor of the Faculty of Intelligent Computing, Tianjin University. His main research interests are deep reinforcement learning and multi-agent systems. He has published more than 100 international conference and journal papers in the field of artificial intelligence, and 2 monographs. He has presided over more than 10 projects such as the 2030 major artificial intelligence project of the Ministry of Science and Technology, the major artificial intelligence cultivation project of the National Science and Technology Foundation Committee, and the key project of national defense science and technology innovation, and his research results have won the best paper award of international conferences for 3 times and the champion of the NeurIPS20-22 conference competition for 4 times. The relevant achievements have been applied in the fields of industrial basic software intelligence, autonomous driving, game AI, advertising and recommendation, 5G optimization, logistics scheduling and other fields.
Title: Reinforcement Learning in the Era of Large Models
Abstract: This report will first introduce the background of traditional reinforcement learning, and then introduce the new learning paradigm of decision-making models in the era of large models, as well as how reinforcement learning can help decision-making models and the challenges and solutions they face, and introduce their applications in scenarios such as autonomous driving, EDA chip design, and embodied intelligence.
About CNCC2024
CNCC2024 will be held on October 24-26 in Hengdian Town, Dongyang City, Zhejiang Province, with the theme of "Developing New Quality Productivity, Computing Leads the Future". The three-day conference included 18 invited reports, 3 conference forums, 138 thematic forums, 34 thematic activities and more than 100 exhibitions. More than 800 speakers, including Turing Award winners, academicians of the Chinese Academy of Sciences and the Chinese Academy of Sciences, top scholars at home and abroad, and well-known entrepreneurs, looked forward to cutting-edge trends and shared their innovative achievements. More than 10,000 people are expected to attend.