ACM SIGAI CHINA aims to carry out academic/technological exchanges on the professional contents of robotics, NLP, visual, big data, cognitive, brain science, and machine learning, and promote the integration of academia and industry closely and the industrial landing of AI technology. SIGAI CHINA will be carry out academic exchanges with ACM China to enhance the influence in national scientific and technological activities and international academic influence.
Lorine Lin (DMAI,Sun Yat-sen University)
Wangmeng Zuo (Harbin Institute of Technology)
Qilong Wang (Tianjin University)
Man Yuan (Beijing Institute for General Artificial Intelligence)
Xiaodan Liang (Sun Yat-sen University)
Guanbin Li (Sun Yat-sen University)
Wenhua Xia (Peking University)
Ke Liu (DMAI)
|13:30-14:00||Openning Remarks||Le Dong, Chair of ACM SIGAI CHINA, Executive Vice President of Beijing Institute for General Artificial Intelligence, Leading and Innovative Talents of Guangzhou，HK Talents Admission Scheme||Guest: Song-Chun Zhu, President of BIGAI, Chair Professor of Tsinghua and PKU|
|14:00-14:35||Keynote 1||Robust Pattern Recognition in Open World||Lorine Lin, Serectary General of ACM SIGAI China, Professor of Sun Yat-sen University, IET Fellow||Cheng-Lin Liu, the Director and Professor of the National Laboratory of Pattern Recognition, Institute of Automation of Chinese Academy of Sciences, Professor at the School of Artificial Intelligence, University of Chinese Academy of Sciences|
|14:35-15:10||Keynote 2||Brain and Cognition Inspired 3D Scene Understanding and Learning||Huimin Ma, Head of Department of Internet of Things and Electronic Engineering of School of Computer&Communication Engineering, Deputy Dean of the Institute of Artificial Intelligence, Vice Chair and Secretary General of China Society of Image and Graphing, Beijing "Woman Pace-Setter"|
(Re-)building Trust in Deep Learning
|Dacheng Tao, Inaugural Director of the JD Explore Academy, Vice President of JD.com， Advisor of the University of Sydney|
|15:45-16:05||Keynote 4||Empowering Industry Innovation with Artificial Intelligence||Tao Mei, IEEE/IAPR Fellow, Vice President with JD.com and the Deputy Managing Director of JD AI Research|
|16:05-16:25||Teabreak (with poster)|
|16:25-17:25||Panel||How to commercialize large-scale pre-trained model?||Guanbin Li, Associate Professor of Sun Yat-sen University, Wuwenjun Excellent Youth Award Winner||
Cheng-Lin Liu, the Director and Professor of the National Laboratory of Pattern Recognition, Institute of Automation of Chinese Academy of Sciences, Professor
at the School of Artificial Intelligence, University of Chinese Academy of Sciences；
Huimin Ma, Vice Chair and Secretary General of China Society of Image and Graphing, Beijing "Woman Pace-Setter" ；
Lorine Lin, Serectary General of ACM China, Professor of Sun Yat-sen University, IET Fellow
Xiaodan Liang, Associate Professor of Sun Yat-sen University，Alibaba DAMO Academy Young Fellow
Wangmeng Zuo, Harbin Institute of Technology
|13:30-13:35||Openning Remarks||Xiaodan Liang, Associate Professor of Sun Yat-sen University||-|
|13:35-14:10||Keynote 1||复杂系统视角下的群体智能理论与技术发展||Wenjun Wu, Professor of Beihang University|
|14:10-14:30||Keynote 2||AI Framework Development Trend and MindSpore Practice||Yu Fan, Senior Architect of Huawei's all-scenario AI computing framework MindSpore，2020 OSCAR Award Winner|
|14:30-14:50||Keynote 3||人工智能赋能金融行业||Liang Xu，平安金融壹账通人工智能研究院总工程师,麻省理工科技评论“全球35岁以下科技创新35人|
|15:10-15:30||Keynote 4||Xiaodan Liang, Associate Professor of Sun Yat-sen University||DMAI Builds New Applications of AI in Basic Education||Rumin Zhang, Postdoc Fellows, HKU,Director of Technology，DMAI|
|15:30-15:50||Keynote 5||When Self-supervised Learning Meets Reasoning||Lorine Lin, Serectary General of ACM SIGAI China, Professor of Sun Yat-sen University, IET Fellow||Xiaodan Liang, Associate Professor of Sun Yat-sen University|
Keynote: Robust Pattern Recognition in Open World
National Laboratory of Pattern Recognition (NLPR)
Institute of Automation of Chinese Academy of Sciences
Abstract: Existing pattern recognition methods mostly concern the classification accuracy, while ignore the rejection and robustness in open world. In recent years, deep learning methods achieved huge successes in pattern recognition, but the popular deep neural networks show poor robustness to noise and outlier in open world. In this talk, I first explain the robustness of pattern recognition, list the related research issues of robust pattern recognition, and introduce some methods for improving the robustness for open set recognition, which needs to both classify within-class patterns and reject outlier. I will give the formulations of two rejection modes （ambiguity rejection and outlier rejection）and introduce some methods. Particularly, I will introduce a newly proposed deep learning method for open set recognition: convolutional prototype network (CPN). The CPN uses a prototype classifier for classification, and learns convolutional feature space and prototypes jointly to yield high accuracy for both classification and outlier rejection. The CPN also shows potential in continual learning and adversarial robustness.
BIO: Cheng-Lin Liu is a Professor at the National Laboratory of Pattern Recognition, Institute of Automation of Chinese Academy of Sciences, and now the Director of the Laboratory. He received the PhD degree in pattern recognition and intelligent control from the Chinese Academy of Sciences, Beijing, China, in 1995. His research interests include pattern recognition, image processing, neural networks, machine learning, and especially the applications to document analysis and recognition. He has published over 300 technical papers at journals and conferences. He is a fellow of the CAAI, the IAPR and the IEEE.
Keynote: brain and cognition inspired 3D scene understanding and learning
Vice Chair and Secretary General of China Society of Image and Graphing
Beijing "Woman Pace-Setter"
Abstract: 计算机视觉与脑科学、认知科学有着紧密的联系，但人类视觉感知模式与计算视觉计算方法之间目前存在着难以逾越的鸿沟，尤其是在人类视觉的量化表征、认知学习计算架构等关键问题上。本报告从脑启发和认知启发两个源头介绍我们在以原型记忆、格式塔完型认知为核心的复杂环境理解、三维目标检测、驾驶行为行为预测等方面的研究。报告围绕智能驾驶任务，介绍三个方面的内容：（1）弱监督、小样本视觉认知学习；（2）Thinking in 3D驾驶场景三维目标检测；（3）Predictive-Bi-LSTM-CRF驾驶行为认知预测。
BIO: 马惠敏教授，博士生导师，2001年博士毕业后在清华大学电子工程系承担教学科研工作，担任三维图像认知与仿真实验室（3DImageLab）负责人，2019年担任北京科技大学计算机与通信工程学院物联网与电子工程系主任、人工智能研究院副院长，现任中国图象图形学学会副理事长兼秘书长，北京市“三八红旗奖章”获得者。 从事计算机视觉与认知心理学交叉研究，探索复杂环境无人系统视觉感认知与决策关键技术。首次在国际上建立了图像认知心理测评智能系统，2016年获得吴文俊人工智能科技创新一等奖，教育部鉴定为“原始性创新，达到国际领先水平”；提出的基于GPU的高效能复杂环境仿真方法及应用，2017年获得教育部技术发明奖二等奖；提出的复杂环境中目标认知方法，2015-2017年连续在国际最大的自动驾驶数据集（KITTI）评测中获得第一名，2018-2019年在驾驶员状态预测国际数据集（Brain4Cars）上获得最好的成绩，2020年获得中国图象图形学学会技术发明一等奖。作为通讯作者在TPAMI、TIP、TITS、PR、CVPR、NIPS、ICCV、ICIP等发表论文100余篇.作为负责人承担了国家重点研发计划子课题、国家自然科学基金、专项重点基金、国际国内企业合作等30余项科研项目，获批及申请专利十余项，两项专利完成了科研成果转化。
Keynote:(Re-)building Trust in Deep Learning
Director of the JD Explore Academy
Vice President of JD.com
Fellow of the Australian Academy of Science, AAAS, ACM and IEEE.
Abstract: The world is in the eve of the enthusiasm revolution by deep learning sweeping across almost all sectors of our society. Concerns are rising when the deployment has happened in the security critic domains, including autonomous vehicles and medical diagnosis. Fatal disasters on road, infamous privacy breaches, and shocking discrimination scandals undermine public confidence in deep learning applications. In this talk, we will present our perspectives, theory, and practice in (re-)building trust in deep learning.
BIO: Dacheng Tao is the Inaugural Director of the JD Explore Academy and a Vice President of JD.com. He is also an advisor of the digital science institute in the University of Sydney. He mainly applies statistics and mathematics to artificial intelligence and data science, and his research is detailed in one monograph and over 200 publications in prestigious journals and proceedings at leading conferences. He received the 2015/2020 Australian Eureka Prize, the 2018 IEEE ICDM Research Contributions Award, and the 2021 IEEE Computer Society McCluskey Technical Achievement Award. He is a fellow of the Australian Academy of Science, AAAS, ACM and IEEE.
Professor of Beihang University
Keynote: AI Framework Development Trend and MindSpore Practice
Senior Architect of Huawei's all-scenario AI computing framework MindSpore
2020 OSCAR Award Winner
Abstract: AI model training requires advanced skills and experts. Difficulties such as the high-tech threshold, high development costs, and long deployment period hinder the development of the AI developer ecosystem in the entire industry. To help developers and the industry better cope with this system-level challenge, the next-generation AI framework MindSpore focuses on simple programming, device-cloud synergy, easy debugging, and excellent performance, reducing the AI development difficulty. In this speech, we will share the AI framework challenges, MindSpore solutions and practices. We are looking forward to a collision of ideas with you.
BIO: Fan Yu, PhD in Computer Science, was graduated from University of Science and Technology of China. He has worked in Huawei for 11 years, and now he is the senior architect of Huawei's all-scenario AI computing framework MindSpore, leading the design and development of architectures and algorithms such as Huawei AI framework, cloud computing resource scheduling, and SDN large-scale routing. He is an excellent open-source developer awarded at 2020 OSCAR and a visiting professor at Harbin Institute of Technology, having published more than 20 papers at top distributed system conferences and AI conferences and more than 30 patents.
Keynote: When Self-supervised Learning Meets Reasoning
Associate Professor, Sun Yat-sen University
Alibaba DAMO Academy Young Fellow
Abstract: Recently, a variety of self-supervised learning and self-training techniques have revolutionized the perception tasks in computer vision and natural language understanding fields, showing the great advances and generalization in unsupervised and cross-modal pretraining. On the other hand, high-level cross-modal reasoning is a long-standing goal in AI field. This talk will present our recent attempts in bridging the self-supervised learning with high-level reasoning task from three aspects. First, a self-supervised network architecture approach named as BossNAS is presented to search superior transformer architecture from scratch, outperforming current ViT, Deit and T2T. Second, to resolve the geometric math problem solving, a critical cross-modality reasoning problem in smart education, we show how the self-supervised learning can be beneficial for extracting high-level features from the geometric diagram, and facilitate in producing explainable program paths for solving math problems. Finally, we present two newly released large-scale datasets for benchmarking the self-supervised, semi-supervised and domain adaptions techniques from 2D object detection and 3D point cloud detection in autonomous driving domain, hosted in ICCV 2021 Workshop&Challenge. SODA10M, containing 10 million unlabeled images and ONCE consists of 1 million LiDAR scenes and 7 million corresponding camera images. We reproduce and evaluate a variety of self-supervised and semi-supervised methods on two datasets.
BIO: Xiaodan Liang is currently an Associate Professor at Sun Yat-sen University. She was a Project Scientist at Carnegie Mellon University, working with Prof. Eric Xing. She focuses on interpretable and cognitive intelligence and its applications on large-scale visual recognition, automatic machine learning and cross-modality dialogue systems. She has published over 80 cutting-edge papers which have appeared in the most prestigious journals and conferences in the field, Google Citation 10000+. She serves as Area Chairs of ICCV 2019, CVPR 2020, NeurIPS 2021, WACV 2021 and Tutorial Chair (Organization committee) of CVPR 2021. She has been awarded the ACM China and CCF Best Doctoral Dissertation Award, the Alibaba DAMO Academy Young Fellow (Top10 under 35 in China), and the ACL 2019 Best Demo paper nomination. She is named one of the young innovators 30 under 30 by Forbes (China). She and her collaborators have also published the largest human parsing dataset to advance the research on human understanding and successfully organized four workshops and challenges on CVPR 2017, CVPR 2018, CVPR 2019, CVPR 2020. She also organized ICML 2019 and ICLR 2021 workshop. Her current research focuses on self-supervised and life-long learning techniques for large-scale task-driven visual understanding.
Keynote:DMAI Builds New Applications of AI in Basic Education
Director of Technology，DMAI
BIO: 张汝民，北京航空航天大学与大唐电信集团联合培养博士，香港大学博士后，爱尔兰里默克大学香农&马可尼研究中心首席研究员，现任暗物智能科技技术总监， 负责公司谙心助教产品的技术和算法研发，在电子与计算机领域国际学术会议和期刊发表论文20余篇，申请国际与国家发明专利30余项。
Keynote: From Technologies to Applications: Empowering Intelligent Supply Chain with Artificial Intelligence
Vice President with JD.com
Deputy Managing Director of JD AI Research
Abstract: With the rapid development of the global economy, an efficient and collaborative supply chain has shown an indispensable position in the global marketization. The supply chain system has developed from the traditional stage to the intelligent supply chain (ISC) stage, which has become a crucial factor to promote industrial upgrading, especially the transformation of the real economy. As the next-generation information technology, artificial intelligence (AI) technology can combine big data, IOT, and blockchain to build the technology foundation of ISC, which guarantees its agility, collaboration, and demand creation. Together with our partners, we are building the National Artificial Intelligence Open and Innovation Platform for Intelligent Supply Chain. This platform will provide open-source toolkits and open services to the whole industry. This report will introduce the latest research progress of this platform, as well as the newest technical products. Particularly, this report will present how to empower the ISC with AI technology, especially how to connect and realize the essential stages of the ISC, i.e., production, logistics, and services.
BIO: Tao Mei is a Vice President with JD.com and the Deputy Managing Director of JD AI Research. Prior to joining JD.com in 2018, he was a Senior Research Manager with Microsoft Research Asia. He has authored or co-authored over 200 publications (with 12 best paper awards) in journals and conferences. He is or has been an Editorial Board Member of leading multimedia journals, and the General/Program Chairs of premier multimedia conferences. He was elected as a Fellow of IEEE (2019), a Fellow of IAPR (2016), and a Distinguished Scientist of ACM (2016), for his contributions to large-scale multimedia analysis and applications.