ICLR 2023 Social Event Proposal:
Safe and Responsible Artificial Intelligence
Topic: Safe and Responsible AI
Time: May 5, 2023 at 01:30 pm , Beijing, Shanghai
Join from PC, Mac, Linux, iOS or Android: https://iso.zoom.us/j/92563929731?pwd=Q0l0RDlCdXVPYXZHb0lGVjR4Vzdidz09  Password: 230505 Or iPhone one-tap : America: +16513728299,,92563929731#  or +16692192599,,92563929731# 

Meeting ID: 925 6392 9731
Password: 230505

INTRODUCTION

In recent months, large language models (LLM) such as ChatGPT and GPT-4 have received widespread attention from the whole society, and the breakthrough has significantly promoted the development of Artificial General Intelligence (AGI). While new technologies bring many new application scenarios, we must carefully consider how the LLMs affect the people who use it and are impacted by it.

Building safe and responsible AI systems is an ongoing area of study. Many relative groups such as government entities, academic and industry are working on the theoretical or practical norms. The area of safe and responsible AI includes many topics such as accuracy, robustness, safety, explainability and interpretability, privacy, reliability, etc.

This social event aims to bring together researchers and practitioners from a variety of communities(e.g., machine learning, computer security, data privacy and ethics) in an effort to synthesize promising ideas and research directions, as well as foster and strengthen crosscommunity collaborations on both theoretical studies and practical applications.

The workshop co-sponsored by the Institute of Artificial Intelligence at Tsinghua University and Alibaba Security, aims to invite top experts and scholars from all over the world to jointly discuss important issues such as cutting-edge AI security technologies, the integration and promotion of AI and security, and AI safety governance.

TIMETABLE

This event will be held in Beijing, with the preliminary location chosen as Tsinghua University and lasting half a day. The tentative schedule is set for 2:00 PM to 6:00 PM Beijing time on May 5th.

This event will revolve around artificial intelligence models, particularly focusing on the safety and credibility of large-scale pre-trained models, and will invite authoritative scholars in the field for discussions. Confirmed invitees include two ICLR Distinguished Paper Award winners and several authoritative experts in the field. Additionally, the forum will invite Chinese ICLR paper authors who cannot attend in person to participate in on-site exchanges, with arrangements for spotlight presentations and poster sessions.

The proposed social event welcomes and encourages diverse audiences and participants, in particular minority and underrepresented groups. All the papers, slides, and posters presented (e.g., keynotes and paper authors) will be made publicly available. Also, together with ICLR conference, we will provide livestream links and record the videos and make them publicly available upon the presenters’ agreement.



Beijing CST(UTC+0800)
Keynote Speeches:
14:00-14:25 Liwei Wang, Professor at Peking University and Adjunct Professor at Tsinghua University's School of Cross Information "Robust Generalization Requires Exponentially Large Models"
14:25-14:50 Yi Zeng, Professor at the Institute of Automation, Chinese Academy of Sciences "From Ethics of Large AI Models to Moral AI"
14:50-15:15 Zhibo Wang, Professor at Zhejiang University's School of Computer Science and Cyber Space Security "Trustworthy AI: Security, Privacy, and Fairness"
15:15-15:40 Xingxing Wei, Associate Professor of the Institute of Artificial Intelligence, Beihang University "Adversarial attacks against intelligent visual models in the physical world"

Young Scholar Presentations:
15:40-15:50 Yinpeng Dong, Postdoctoral Researcher at Tsinghua University's Department of Computer Science "Robustness of Visual Recognition in the 3D World"
15:50-16:00 Jiawen Deng, Postdoctoral Researcher at Tsinghua University's Department of Computer Science "On the Safety of Large Language Models"

Discussion and Exchange:
16:00-16:40 Panel: How to Govern Security and Ethical Issues Brought by AI
16:40-17:25 Paper Spotlight Sharing

Jun Cen , “The Devil is in the Wrongly-classified Samples: Towards Unified Open-set Recognition” Rongjie Huang, “TranSpeech: Speech-to-Speech Translation With Bilateral Perturbation”

Jin Li, “Progressively Compressed Auto-encoder for Self- supervised Learning”

Chang Liu, “Revocable Deep Reinforcement Learning with Affinity Regularization for Outlier-Robust Graph Matching” Changze Lv, “Spiking Convolutional Neural Networks for Text Classification”

Fengda Zhang, “Fairness-aware Contrastive Learning with Partially Annotated Sensitive Attributes”

Zeming Wei, “CFA: Class-wise Calibrated Fair Adversarial Training”

Yunhao Zhang, “Crossformer: Transformer Utilizing Cross- Dimension Dependency for Multivariate Time Series Forecasting”
17:25-18:00 Poster Technical Achievement Exhibition (Tea Break)
18:00 End

We briefly provide the topics of each section as well as the organizers. Additionally, we include the invited speakers with their affiliations. All the organizers and invited speakers have confirmed their participation. Currently, we have 4 confirmed keynote speakers distributed across 4 different sections.

KEYNOTES

● Professor: Liwei Wang (Peking University)

The research field is machine learning theory. He has published over 100 papers in top conferences and journals on machine learning. He has served as the domain chairman of top level machine learning conferences NIPS/NeurIPS, ICML, and ICLR for many times, and currently serves as the editorial board member of IEEE T. Pattern Analysis and Machine Intelligence.

● Professor: Xingxing Wei (Beihang University)

He worked as a postdoctoral researcher in the Department of Computer Science at Tsinghua University from 2017 to 2019. Prior to that, he received his bachelor's degree from Beihang University and his Ph.D. degree from Tianjin University. After graduation, he worked as a senior computer vision algorithm engineer at Alibaba for one year. His research interests mainly lie in adversarial machine learning, computer vision, etc., and he has published more than 60 academic papers in TPAMI, IJCV, NeurIPS, CVPR, ICCV, ECCV, IJCAI, AAAI, ACMMM. His team has won the champion of CAAD CTF of DEFCON2018, the runner-up of the Robust Logo Detection Competition held by ACMMM2021, and the 4th place in the ImageNet Unrestricted Counter Attack Competition held by CVPR2021. As the project leader, he has presided over the "New Generation Artificial Intelligence" 2030 major project topic of the Ministry of Science and Technology, the general project and youth project of the National Natural Science Foundation of China, the CCF-Tencent Rhinoceros Bird Fund, and so on.

● Professor: Yi Zeng (Chinese Academy of Sciences)

The research field mainly involves brain-like artificial intelligence and artificial intelligence ethics. Inspired by the structure and mechanism of the brain, the research work builds a brain like neural network and cognitive computing model, which is applied to brain like intelligent robots; And conduct research on the risks, safety, and ethics of artificial intelligence to ensure that research on brain-like artificial intelligence develops in a direction beneficial to humans and society.

● Professor: Zhibo Wang (Zhejiang University)

Research interests include the Internet of Things, artificial intelligence security, network security, and privacy protection. He has published more than 100 papers in international authoritative journals and academic conferences, including over 30 Class A journals and conference papers recommended by CCF, and seven highly cited papers from ESI. Won the IEEE HPCC 2019 Outstanding Paper Award, Outstanding Scientist and exemplary individual of the Institute of Electronics.

ORGANIZERS

● Hang Su (suhangss@mail.tsinghua.edu.cn): Dr. Hang Su is an associate professor in the Department of Computer Science and Technology at Tsinghua University. Before joining Tsinghua, he received his Ph. D. degree from Shanghai Jiaotong University in 2014 and worked as a visiting scholar at Carnegie Mellon University from 2011 to 2013. His research interests lie in the development of computer vision and machine learning algorithms with a recent focus on robust and interpretable machine learning algorithms. He has published around 50 papers including CVPR, ECCV, TMI, and served as senior PC or PC members in the dominant international conferences including IJCAI, AAAI, CVPR. He received “Young Investigator Award” from MICCAI2012, the “Best Paper Award” in AVSS2012, and “Platinum Best Paper Award” in ICME2018.
Webpage:http://www.suhangss.me/
Google Scholar: https://scholar.google.com/citations?hl=en&user=dxN1_X0AAAAJ

● Yuan He (heyuan.hy@alibaba-inc.com): Dr. Yuan He is the Chief Scientist of Alibaba AI Governance Research Center (AAIG). He received his B.E. and Ph.D. degrees from Tsinghua University. His research areas include machine learning, pattern recognition, computer vision and AI security. Dr. He is the author of more than 60 international journal and conference articles, such as TPAMI, ICML, CVPR, NeurIPS. He is also the inventor of more than 20 patents filed in China, Europe and US. In recent years, Dr. He served as organizer or PC member of workshops and challenges related to adversarial examples and model robustness on several international conferences. He is also the spec editor of IEEE and CCSA AI Robustness standards.
Google Scholar: https://scholar.google.com/citations?user=cWbXLzgAAAAJ&hl=zh-CN

● Jun Zhu (dcszj@tinghua.edu.cn) is a Professor at the Department of Computer Science and Technology in Tsinghua University. He was an Adjunct Faculty at the Machine Learning Department in Carnegie Mellon University from 2015 to 2018. His research interest lies in machine learning and applications in text and image analysis. Dr. Zhu has published over 100 papers in the prestigious conferences and journals. He is an associate editor-in-chief for IEEE Trans. on PAMI. He served as area chair/senior PC for ICML, NeurIPS, IJCAI, UAI, AAAI, and AISTATS. He was a local co-chair of ICML 2014. He is a recipient of several awards, including IEEE Intelligent Systems "AI's 10 to Watch" Award, MIT TR35 China, NSFC Excellent Young Scholar Award, CCF Young Scientist Award, and CCF first-class Natural Science Award. His work is supported by the National Youth Top-notch Talent Support program.
Webpage: https://ml.cs.tsinghua.edu.cn/~jun/index.shtml
Google Scholar: https://scholar.google.com/citations?user=axsP38wAAAAJ&hl=en

● Hui Xue (hui.xueh@alibaba-inc.com): Dr. Hui Xue is a Senior Staff Engineer in Alibaba Group. He received his Ph.D. degreee from the University of Zhejiang University and currently leads the Alibaba Turing Security Lab in Alibaba’s Security Department. Dr. Xue has over 10 years of experience in applying computer vision, NLP and large-scale machine learning methods into solving real-world industrial problems in areas such as cybersecurity, risk management and searching. He and his team has published more than 20 papers at top international conferences and holds a number of patents.




A reminder for all attendees: according to ICLR's official requirements, all participants must register virtually. If you haven't registered, you will not be allowed to enter the venue (website: https://iclr.cc/Register2).

Note: If you are participating in the spotlight, poster presentation, or other aspects of this event, you can apply for registration funding from the organizers. Due to space limitations at the venue, please confirm whether you can attend in person to give more attendees the opportunity to participate.