【2021.05.29】Improving and Securing Machine Learning Systems

报告题目Improving and Securing Machine Learning Systems

主讲嘉宾Dr. Bolun Wang    

报告时间2021年5月29日 14:00-16:00

报告地点复旦大学江湾校区交叉二号楼 A4009


报告摘要Machine Learning has already become a crucial piece of our lives, as it powers a wide range of applications. From image recognition, self-driving cars, to natural language processing, and knowledge extraction, ML has fundamentally transformed every domain it has touched. Despite all the success, a key problem preventing the wider adoption of ML is its non-transparency: ML models appear to us as complex numerical black-boxes and we do not understand how they operate. Despite numerous efforts from the community in recent years, we still see limited progress towards complete understanding of ML. Meanwhile, the implications of non-transparency are imminent and severe. It prevents us from understanding when and where ML systems would fail, which enables adversarial attacks against ML models. More concerning that that, it enables the possibility of hiding malicious behaviors inside ML models, i.e. backdoors. Such backdoors could compromise the ML model and alter its behavior once activated by the attacker.The threats are imminent, putting currently deployed ML models at risk. But the progress towards fully understanding ML models is limited so far, and unlikely to significantly advance in the near future. This conflict between imminent threats and limited progress calls for immediate solutions to address some of the most important issues mentioned above.In this talk, I will cover our work to improve the security of ML as opaque systems. First, I will present a new adversarial attack against DL models we propose in the practical setting. We specifically focus on the context of transfer learning, which is the most promising solution for normal users to train high-quality DL models with limited data. Second, I will discuss backdoor attacks in DL models, where hidden behaviors are injected into DL models, and cause misclassification for any input containing certain trigger. I will present our design to detect and mitigate such hidden backdoors.

关于嘉宾Dr. Bolun Wang is a research scientist at Facebook, working on fighting against scraping on the FB family of apps. He works at the External Data Misuse team under the Privacy pillar. Dr. Wang graduated from Computer Science Department at UC, Santa Barbara. He was working in the SAND Lab (Systems, Algorithms, Networking and Data), co-advised by Professor Ben Y. Zhao and Professor Heather Zheng. He received my bachelor degree of Electronic Engineering from Tsinghua University in 2013. His Ph.D. thesis focuses on the Security and Performance of Deep Learning. It is mainly about identifying vulnerabilities and novel attacks caused by the non-transparent nature of Deep Learning models. His work tries to identify most critical threats and design defenses to secure DL systems. His general research interest is Security and Privacy. Apart from Deep Learning systems, his work also covers Mobile applications, Online Social Networks (OSNs), and Crowdsourcing systems. He is particularly interested in measurement and data-driven analysis of such systems. His goal is to further understand Internet systems, identify security issues, and provide practical solutions.

地址: 中国 上海市杨浦区淞沪路2005号复旦大学江湾校区2号交叉学科楼
邮编: 200438
电话: +86-21-31242153
传真: +86-21-31242153
E-mail: dataology@fudan.edu.cn