The 2024 INTERNATIONAL workshop on
Artificial Intelligence & Cyber Security
Menu
This talk explores recent trends and challenges at the intersection of AI and cybersecurity, focusing on two fundamental questions: How can we protect user inputs and model outputs? How can we defend AI models themselves from attacks?
Professor Hayato Yamana will provide an overview of the latest developments in secure AI, including threats such as data leakage, adversarial manipulation, and model theft. He will also introduce ongoing research from his group at Waseda University, which aims to enhance trust and resilience in AI systems through advanced security and privacy-preserving techniques.
Professor Yamana is a leading expert in AI, cybersecurity, and big data analysis. He received his Doctor of Engineering degree from Waseda University in 1993 and has been a Professor there since 2005. He
currently serves as Vice President for IT Promotion and Chief Information Officer at Waseda University. He has also held major leadership roles in Japan’s information research community, including Director of the Database Society of Japan (DBSJ), Director of the Information Processing Society of Japan (IPSJ), and Vice Chair of the
IEICE Information and Communication Society. Professor Yamana’s work on fully homomorphic encryption and secure data analysis aligns closely with the AICS Workshop’s themes. His deep expertise and leadership perspective would make him an excellent keynote speaker for this year’s event.
The approach and process of Data Poisoning Attacks (DPA) to distort training data to machine learning model and manipulate the model behaviours is not only technically complex but also often victim model dependent. To protect the victim model, the vast number of DPAs and their variants make defenders rely on trial and error techniques to find the ultimate defence solution which is exhausting and very time-consuming. This talk summarises the latest research on DPAs and defences, proposes a DPA characterizing model to help investigate adversary attacks dependency on the victim model, and builds a DPA roadmap as the path navigating to defence. Having the roadmap as an applied framework that contains DPA families sharing the same features and mathematical computations will equip the defenders with a powerful tool to quickly find the ultimate defences, away from the exhausting trial and error methodology. The roadmap validated by use cases has been made available as an open access platform, enabling other researchers to add in new DPAs and update the map continuously.
Dr. Pang is an Associate Professor of cyber security and he leads the Internet Commerce Security Lab (ICSL) at the Institute of Innovation, Science and Sustainability, Federation University Australia. Before joining Federation University, he was a Professor of Data Analytics and Director of Centre Computational Intelligence for Cybersecurity at the Unitec Institute of Technology, New Zealand. His research specializes in AI for cybersecurity, applied blockchain, and digital agricultural traceability. Dr. Pang is a Senior Member of IEEE, the Event Editor of Neural Network Journal Elsevier, Associate Editor for Springer’s Pattern Analysis & Applications, and Advisory Board Member for the Journal of Cybersecurity Technology (Taylor & Francis).
To be announced soon.