主讲人：吴帆
讲座时间：2024-09-27 10:00:00
讲座地点：学术楼105
主办单位：教师工作部/人事处，计算机科学与技术学院
主讲人简介：吴帆博士，现为上海交通大学计算机科学与工程系特聘教授、博导、系主任，2020年国家自然科学基金杰出青年科学基金获得者。在大小模型协同智能、分布式智能系统、大数据管理、智能博弈等领域取得了一系列科研成果，已发表学术论文200余篇，论文发表在JSAC、TON、TMC、TPDS、TOC、TKDE等国际著名期刊，以及MOBICOM、OSDI、INFOCOM、ICDE、VLDB、KDD等重要国际会议，并出版英文专著一部。先后担任IEEE Transactions on Mobile Computing、ACM Transactions on Sensor Networks、IEEE Transactions on Network Science and Engineering、Elsevier Computer Networks编委，以及Journal of Computer Science and Technology青年编委和中国计算机学会通讯专题编委。曾获国家级教学成果奖二等奖、教育部自然科学奖一等奖、上海市科技进步奖一等奖、CCF科技进步奖一等奖、ACM中国新星奖、CCF-IEEE青年科学家奖、仲英青年学者、CCF-腾讯犀牛鸟卓越奖、CCF-Intel青年学者提升计划、上海市浦江人才，以及包括IEEE/ACM IWQoS 2020在内的7次国际学术会议论文奖。先后担任ACM中国理事会副主席、CCF上海分部秘书长、CCF YOCSEF上海主席等。作为项目负责人承担科技创新2030“新一代人工智能”重大项目、国家自然科学基金、上海市科委项目20余项。
Fan Wu is a professor and chair of the Department of Computer Science and Engineering, Shanghai Jiao Tong University. His research interests include wireless networking and mobile computing, data management, algorithmic network economics, and privacy preservation. He has published more than 200 peer-reviewed papers in technical journals and conference proceedings. He is a recipient of the first class prize for Natural Science Award of China Ministry of Education, China National Fund for Distinguished Young Scientists, ACM China Rising Star Award, CCF-IEEE CS Young Computer Scientist Award, CCF-Tencent ``Rhinoceros bird'' Outstanding Award, and CCF-Intel Young Faculty Researcher Program Award. He has served as an associate editor of IEEE Transactions on Mobile Computing, IEEE Transactions on Network Science and Engineering, and ACM Transactions on Sensor Networks, an area editor of Elsevier Computer Networks, and as the member of technical program committees of more than 100 academic conferences.
讲座内容：人工智能的快速发展与广泛应用已成为经济社会发展的强大引擎。大小模型端云协同进化作为人工智能关键前沿技术，可有效克服云侧集中式学习范式在可扩展性、实时性、个性化、负载成本、隐私安全等方面的不足，已成为产学研的焦点，并被Gartner和阿里巴巴达摩院等国内外知名机构预测为革新智能计算范式的科技趋势。本报告将会追寻端云协同智能技术的发展脉络，分享我们团队在端侧智能推理、大规模联合学习以及端云协同分布式智能支撑系统等方面的研究进展。
The rapid development and widespread application of artificial intelligence have become a powerful engine for economic and social development. As a key cutting-edge technology in artificial intelligence, the collaborative evolution of heterogeneous models can effectively overcome the shortcomings of cloud side centralized learning paradigms, in terms of scalability, real-time performance, personalization, cost, and privacy preservation. It has attracted the attention of both industry and academia, and has been recognized as a new paradigm of intelligent computing by both Gartner and Alibaba's Barbados Institute. In this talk, I will share our recent progress on large-scale distributed learning with heterogeneous models.
