Keynote: Computer Science and AI Are Driving an Information Revolution

John Hopcroft, winner of the A. M. Turing Award in 1986, IBM Professor of Engineering and Applied Mathematics, Department of Computer Science, Cornell University. He received both his M.S. (1962) and Ph.D. (1964) in electrical engineering from Stanford University. After three years of faculty at Princeton University, he joined Cornell College and was named a professor in 1972. Since 1985, he has been the Joseph C. Ford Professor of Computer Science. The Association for Computing Machinery (ACM) awarded him the A. M. Turing Award in 1986 in recognition of his fundamental research in algorithms and data structures. Professor Hopcroft is an academician of the American Academy of Sciences, the American Academy of Engineering and the American Academy of Arts and Sciences, American Association for the Advancement of Science Fellow, AAAS Fellow, IEEE Fellow and ACM Fellow. In 1992, he was appointed by President Bush to the National Science Board (NSB), which oversees the National Science Foundation (NSF), and served through May 1998. Professor Hopcroft is also a member of the Scientific Advisory Committee of the David and Lucy Packard Scholarship in the field of science and engineering, the Financial Management Committee of the Industrial and Applied Mathematics Society, the New Delhi Advisory Board, the Microsoft's technical Advisory Board for research Asia, and the Engineering Advisory Board, Seattle University, and the Chilean Millennium Science Project Committee.

Keynote: Challenges to Computing: The Age of Digital Transformation

BIO:梅宏，中国科学院院士，发展中国家科学院院士，欧洲科学院外籍院士， IEEE Fellow，中国计算机学会（CCF）理事长。主要从事软件工程和系统软件领域研究，曾荣获国家技术发明一等奖、何梁何利基金科学技术进步奖等重要奖项。历任国家863计划专家组组长，国家“核高基”科技重大专项专家组成员，全国信息技术标准化技术委员会大数据标准工作组组长，国家重点科技研发专项“云计算和大数据”实施方案编制组组长、总体组组长等。

Keynote: 通用人工智能：学科大交叉大融合的机遇与挑战

Abstract:This talk will present the recent development and trend of artificial intelligence as a scientific discipline, from the perspective of multi-disciplinary integration and unification, and discuss the current interest in returning to pursuing general intelligence which is becoming a focus in global competition for the next 10-20 years. Constructing general AI agents is an extremely complex and daunting task. To reach this goal, we may need to exam our current research methodology and paradigm, explore completely new ways for organizing AI research, and build AI as a new independent discipline with its own core areas and innovative curriculum. In the end, the talk will discuss the potential impacts of AI on digital economy and social governance.

BIO: 朱松纯教授，出生于湖北省鄂州，全球著名计算机视觉专家，统计与应用数学家、人工智能专家。 1991年毕业于中国科技大学，1992年赴美留学，1996年获得美国哈佛大学计算机博士学位。后来在布朗大学、斯坦福大学、俄亥俄州立大学等地工作，2002年至2020年，在美国加州大学洛杉矶分校（UCLA）担任统计系与计算机系教授，UCLA视觉、认知、学习与自主机器人中心主任。在国际顶级期刊和会议上发表论文300余篇，获得计算机视觉、模式识别、认知科学领域多个国际奖项， 包括3次问鼎计算机视觉领域国际最高奖项--马尔奖，赫尔姆霍茨奖等，2次担任国际计算机视觉与模式识别大会主席（CVPR2012、CVPR2019），2010-2020年 2次担任美国视觉、认知科学、人工智能等领域多大学、跨学科合作项目MURI负责人。朱松纯教授长期致力于构建计算机视觉、认知科学、乃至人工智能科学的统一数理框架。
在留美28年后， 朱教授于2020年9月回国，组建北京通用人工智能研究院，并任北京大学讲席教授、北京大学人工智能研究院院长、清华大学基础科学讲席教授等职务。

Keynote: A new roadmap for linking theories of programming

Abstract:Formal methods advocate the crucial role played by the algebraic approach in specification and implementation of programs. Traditionally, a top-down approach (with denotational model as its origin) links the algebra of programs with the denotational representation by establishment of the soundness and completeness of the algebra against the given model, while a bottom-up approach (a journey started from operational model) introduces a variety of bisimulations to establish the equivalence relation among programs, and then presents a set of algebraic laws in support of program analysis and verification. This talk proposes a new roadmap for linking theories of programming. Our approach takes an algebra of programs as its foundation, and generates both denotational and operational representations from the algebraic refinement relation. This talk demonstrates the application of this new approach to GCL (Guarded Command Language) and CSP (Communicating Sequential Processes) to link their various semantical representations based on their algebraic semantics.

BIO:He Jifeng is a scientist of computer software, academician of Chinese Academy of Sciences. He is the first president of the Software Engineering Institute, East China Normal University. He is the founder of Unifying Theories of Programming, the founder of complete theory of data refinement, the pioneer of the theory and technology on trustworthy software design, and the advocator of the trustworthy artificial intelligence. In recent years, His main research focuses on the key technology of trustworthy artificial intelligence to ensure the safety and reliability of national artificial intelligence applications.

Keynote: A Mathematical Perspective of Machine Learning

Abstract:The heart of modern machine learning (ML) is the approximation of high dimensional functions. Traditional approaches, such as approximation by piecewise polynomials, wavelets, or other linear combinations of fixed basis functions, suffer from the curse of dimensionality (CoD). This does not seem to be the case for the neural network-based ML models. To quantify this, we need to develop the corresponding mathematical framework.In this talk, I will report the progress made so far and the main remaining issues within the scope of supervised learning. I will discuss three major issues: approximation theory and error analysis of modern ML models, qualitative behavior of gradient descent algorithms, and ML from a continuous viewpoint.

BIO:Weinan E received his Ph.D. from UCLA in 1989. After being a visiting member at the Courant Institute of NYU and the Institute for Advanced Study at Princeton, he joined the faculty at NYU in 1994. He is now a professor of mathematics at Princeton University, a position he has held since 1999. Weinan E's work centers around multi-scale modeling and machine learning. He is a pioneer in the area of integrating machine learning and physical modeling to solve problems in traditional areas of science and engineering, such as molecular dynamics, PDEs, and control theory. Weinan E is the recipient of the ICIAM Collatz Prize, SIAM R. E. Kleinman Prize, von Karman Prize, the SIAM-ETH Peter Henrici Prize, and the ACM Gordon-Bell Prize. He is a member of the Chinese Academy of Sciences, a fellow of the American Mathematical Society, a SIAM fellow and a fellow of the Institute of Physics.

Keynote: Computational Intelligence in Music Art

Abstract:音乐曾经是数学的分支。虽然几百年来，艺术与科学走过各自不同的发展道路，很多著名艺术家和科学家深信艺术与科学密切相关，相互影响。
艺术形象思维启发科学创新，科学技术进步推动艺术的发展。对著名的“李约瑟命题”和“钱学森之问”的探讨，既适用于科学也适用于艺术。艺术与科学的交汇促进理工与艺术教育的共同发展。
隐藏在优美旋律中的数学物理规律，竟然同众多自然、工程和社会系统中的规律一致，能够定量分析，对艺术创作产生重要影响。
报告讨论音乐旋律的三个数学特征，由此建立数学模型，揭示作曲家追求旋律变化的有约束熵最大，从而求解得到音乐旋律变化的幂律。研究结果有助于深度分析音乐艺术特别是作曲理论中的计算智能，探索人工智能辅助作曲的定量化方法。

BIO:Xiaohong Guan, academician of the Chinese Academy of Sciences, system engineer, dean of the Faculty of Electronic and Information Engineering at Xi'an Jiaotong University, chief scientist of the Ministry of Education Key Lab For Intelligent Networks and Network Security, Director of Center for Intelligent and Networked Systems at Tsinghua University. Xiaohong Guan received a bachelor's degree and a master's degree in engineering at Tsinghua University in 1982 and 1985 respectively, and a doctorate degree at the University of Connecticut in 1993. He visited Harvard University from 1999 to 2000 and is served as a senior consultant engineer for PG&E in 1993 to 1995. Since 1995, he has served successively as professor, director of the Institute of Systems Engineering, and dean of the Faculty of Electronic and Information Engineering at Xi’an Jiaotong University. From 1999 to 2009, he served as the director of the State Key Laboratory of Manufacturing Systems Engineering, awarded the National Outstanding Youth Fund in 1997, served successively since 2001 as Member of the Chair Professor Group of Tsinghua University, double-employed professor, Director of Center for Intelligent and Networked Systems of Department of Automation, served as the director of the Department of Automation at Tsinghua University from 2003 to 2008, and was elected as an academician of the Chinese Academy of Sciences in 2017. Guan develops the theory of optimization of discrete and mixed manufacturing systems, and proposes a new method of power supply resource optimization bidding strategy and power purchase optimization allocation, and the "opportunistic collusion" bidding game theory of multiple Nash equilibrium points. He leads a multidisciplinary research team, combines systematic scientific methods with network technology to research and develop integrated network security defense systems.Guan has published more than 180 articles. He was awarded the 1996 Li Foundation Heritage Prize, 2008 IEEE Communications Society System Integration and Modeling Best Paper Award, 2004 He Pan Qingyi Discrete Event Dynamic System Best Paper Award, the second prize of 2005 National Natural Science, the second prize of 2006 National Science and Technology Progress Award. He was elected IEEE Fellow in 2006, and Distinguished Lecturer of IEEE Robotics and Automation Society in 2008.

Keynote: Think Sequential, Run Parallel

Abstract: This talk tackles two issues in connection with parallel graph computations. (1) Is it possible to simplify parallel programming, from think parallel to think sequential? That is, we want a parallel system such that we can plug in sequential graph algorithms, and the system parallelizes computations across a cluster of machines, without degradation in performance or functionality of existing graph query engines. (2) Does there exist a parallel model that optimizes computation by adaptively switching between BSP (Bulk Synchronous Parallel) and AP (Asynchronous Parallel) models? That is, the model retains the advantages of BSP and AP, while it reduces stragglers and redundant stale computations inherent to BSP and AP. We answer both questions in the affirmative.

BIO:Professor Wenfei Fan is the Chair of Web Data Management at the University of Edinburgh, UK, and the Chief Scientist of Shenzhen Institute of Computing Science, China. He is a Foreign Member of Chinese Academy of Sciences, a Fellow of the Royal Society (FRS), a Fellow of the Royal Society of Edinburgh (FRSE), a Member of Academia Europaea, and an ACM Fellow. He received his PhD from the University of Pennsylvania (USA), and his MSc and BSc from Peking University (China). He is a recipient of Royal Society Wolfson Research Merit Award (2018), ERC Advanced Fellowship (2015), the Roger Needham Award (2008, UK), Yangtze River Scholar (2007, China), the Outstanding Overseas Young Scholar Award (2003, China), the Career Award (2001, USA), and several Test-of-Time and Best Paper Awards (Alberto O. Mendelzon Test-of-Time Award of ACM PODS 2015 and 2010, Best Paper Awards of SIGMOD 2017, VLDB 2010, ICDE 2007 and Computer Networks 2002). His current research interests include database theory and systems, in particular big data, data quality, data sharing, distributed query processing, query languages, recommender systems and social media marketing.

Keynote: Scale Realization of AI: Exploration and Reflection

Abstract:The technology and application of artificial intelligence are advancing rapidly in recent years, but there are still many difficulties in its large-scale realization in practice. This talk attempts to analyze the pattern changes of developing and applying artificial intelligence technology in the real-world. And, combining with many years of AI experiences in smart city, manufacturing, Internet, health care, education, and agriculture, including technology innovation, product building and application landing, we discuss the challenges, possible path and trends of today and future's large-scale realization of artificial intelligence.

BIO:Xian-Sheng Hua is now a Distinguished Engineer/VP of Alibaba Group, Head of AI Center and City Brain Lab of DAMO Academy, leading a team working on large-scale visual intelligence on the cloud. Dr. Hua is an IEEE Fellow, and ACM Distinguished Scientist. He has authored or coauthored more than 200 research papers and has more than 60 granted patents. He was one of the recipients of the 2008 MIT Technology Review TR35 Young Innovator Award for his outstanding contributions on video search. Hua served as a general co-chair of ACM Multimedia 2020.

Keynote: High Confidence Software Technologies and Our Practice

Abstract:With the increasing demand for high-confidence software in safety-critical fields such as aerospace, national defense, nuclear power, finance, medical care, transportation, and intelligent driving, the research on high-confidence software technology has attracted more and more attention. How to turn these research results into usable and effective tools has become a problem that people are concerned about. This report will introduce our thinking, practice and future challenges based on the development of high-confidence software tools from the Guochuang Software Co.,Ltd.

BIO:Doctor Ji Jinlong, joined GuoChuang Software Co., Ltd. in 2015. He participated in the establishment of Anhui USTCHCS-Guochuang High-Confidence Software Co.,Ltd. in 2017 and lead the product research and development of High-Confidence Software’s Process Analysis Toolkit. At present, Dr.Ji is the director of GuoChuang Software Co., Ltd. , the president of GuoChuang Central Research Institute, and the executive deputy general manager of High-Confidence Software Co., Ltd.

Keynote: The Industrial Applications of Artificial Intelligence

Abstract:Artificial Intelligence is the key driver of the new round of technological revolution and industrial innovation. Industries are demanding for applications of artificial intelligence, and AI is bringing efficiency and new value to various scenarios in every stage which determines the economic benefits of enterprises.
At present, AI-Industries integration innovation is stepping into a new stage. During the process of integrating with industries，Baidu produces Scenario-oriented AI products and solutions to solve the actual problems of its customers. Also, Baidu continues to lower the level of difficulty in developing AI applications, accelerating the evolution of AI technologes, and continuing to build Baidu AI platform, to better support the industrial transformation and upgrading.

BIO:Tian Wu, Baidu Corporate Vice President, Deputy Director of National Engineering Laboratory for Deep Learning Technology and Applications. She is the head of Baidu AI Technology Platform and Baidu Cloud AI Products, overseeing Natural Language Processing, Knowledge Graphs, Computer Vision, Augmented Reality, Big Data Technology, PaddlePaddle, Baidu Cloud AI Products and Baidu Input Method Product.
Tian Wu joined Baidu in 2006, she has been continued to accumulate in the field of artificial intelligence over the years. Under her supervision, AI Technology group is making efforts in developing cutting-edge technologies such as Semantic Understanding, Machine Reading Comprehension, Dialogue System, Machine Translation, Knowledge Graph, Computer Vision, Augmented Reality, etc. PaddlePaddle is the first open-source deep learning platform fully developed and widely used in China. Baidu Brain is the biggest AI open platform in China, and has become an important infrastructure for industrial intelligence. She has been awarded the First Prize of Scientific and Technological Progress Award of Chinese Institute of Electronics four times.
Tian Wu was highlighted in 2020《Fortune》global 40 under 40 technology category list .

Keynote: Towards Compute-Native Networking

Abstract: In this talk, we advocate a new paradigm of interconnection, named compute-native networking. The motivation of compute-native networking is to satisfy the ever increasing demand for computing driven by big data and AI applications. With the slow-down of Moore’s law and the end of Dennard scaling, we need to parallelize this excessive amount of computing among massively scale-out networked clusters with heterogeneous computing resources. But existing interconnection technologies fall short to meet the requirements in performance, latency and scalability. Compute-native networking proposes a unified, scalable memory semantic interconnection to efficiently connect tens-of-thousands computing components with diverse domain specific architectures (DSAs) like CPU, GPU, NPU, memory and storage. This talk will give an overview of an initiation in Huawei towards compute-native networking, called UB. I will go over UB’s design principles, its current progress and potential applications. We believe compute-native networking is a paradigm shift and would shape the future distributed computer systems and applications.

BIO: Kun Tan is Vice President of CSI, heading the Distributed and Parallel Software Lab, Huawei. In past, he has been working on various aspects in networking and networked systems, AI/Serverless frameworks as well as cloud computing. Before joining Huawei, he was a Senior Researcher/Research Manager in Microsoft Research Asia. He has published over 100 papers in top conferences and journals. He received USENIX Test-of-Time Award in 2019.