Address

Contact Information

Email: ycai0031 AT uni DOT sydney DOT edu DOT au

Yipeng Cai

Master Student

Biography

Yipeng Cai is a Master of Computer Science (Data Science and AI Specialisation) student at the University of Sydney. He acquired his double bachelor’s degree of electronic business and economics in Zhong Nan University of Economics and Laws in 2015. He has 9 years working experiences. He started from being a demand engineer and as a management trainee in UG group in 2016. After that, he consistently worked in management positions, successively serving as the group’s planning manager and the compliance and risk control head of its investment company. Since 2020, he participated in two startup projects: a cross-border e-commerce company and an information technology integration company. He has participated in several smart city and smart campus projects, including provincial energy management systems and campus security management systems. His current areas of interest include the Internet of Things (IoT), crowdsourced IoT services, trusted data, trusted data tampering detection, and AI-tampering.

Project Title:

Detection of AI-assisted Tampering of Trust Data in Crowdsourced IoT Services

Project Description

 With the development of the Internet of Things (IoT), the boundaries between people are constantly being broken down. People are becoming more interdependent and connected, this brings unprecedented opportunities for the development of crowdsourced IoT services. However, crowdsourced IoT services carry certain trust risks. These services typically require the participation of numerous third parties to collect trust records. Because trust data uses a distributed storage model, and these third parties may have competitive relationships, the trust data can be tampered with locally to maximize the interests of the third parties. In the context of AI, such tampering becomes extremely easy. We will simulate a crowdsourced IoT service scenario to create a dataset of AI-tampered trust records. Furthermore, we will build our detection algorithm based on subtle clues of AI-tampering.