About Me
Hi 👋, I am Shiyi Yang (Angela), a Postdoctoral Research Fellow in the School of Computer Science and Engineering at the University of New South Wales (UNSW Sydney), working with Prof. Lina Yao. Before that, I received my PhD degree in Computer Science from UNSW Sydney, supported by Commonwealth Scientific and Industrial Research Organisation (CSIRO)’s Data61, Australia, in 2026, under the supervision of Prof. Lina Yao, Dr. Chen Wang, Dr. Xiwei Xu, and Prof. Liming Zhu. I also received my MPhil degree in Computer Science from UNSW Sydney in 2022 under the supervision of Dr. Hui Guo and Prof. Jingling Xue.
My research focuses on trustworthy AI, with an emphasis on AI safety, security, robustness, and alignment in intelligent systems. In particular, I study recommender systems and LLM-based agents, including GUI agents and embodied agents. My work has been published in top-tier venues including WWW, KDD, CIKM, and ICDM. I also serve as a reviewer and PC member for leading conferences and journals, including WWW, KDD, SIGIR, TOIS, and TKDE. I am always open to collaborations and discussions on trustworthy AI, recommender systems, and LLM-based agents.
Research Vision
My research is driven by a fundamental question: how can we build AI systems that remain trustworthy, robust, and aligned in complex and dynamic environments?
My research journey began during my MPhil in cyber security, focusing on deep learning for anomaly detection and threat analysis across IoT, edge, and cloud computing systems. A key lesson from this experience is that effective defenses require a deep understanding of attacks and system vulnerabilities. During my PhD, I extended this perspective to recommender systems, investigating emerging attack surfaces and developing defenses to improve the robustness and trustworthiness of intelligent systems. More recently, I have expanded this line of research to LLM agents, focusing on AI safety & security, robustness, and alignment in increasingly autonomous environments.
Looking forward, I aim to develop trustworthy AI systems that can reliably operate in the real world, remain resilient to adversarial influences, and stay aligned with human values and intentions.
News
- 2026.08: 🎉 One paper accepted to TIST!
- 2026.06: 🎉 Invited to serve as a Session Chair at ICDH 2026!
- 2026.06: 🎉 Graduated with a PhD degree from UNSW and CSIRO!
- 2026.05: 🎉 Started as a PostDoc at UNSW!
- 2026.05: 🎉 One paper accepted to KDD 2026!
- 2026.01: 🎉 One paper accepted to WWW 2026!
- 2025.06: 🎉 Recognized as an Outstanding Reviewer by ACM SIGKDD 2025!
- 2025.08: 🎉 Our work on poisoning recommender systems appeared on ArXiv!
- 2025.04: 🎉 Received the UNSW Industry Engagement Grant!
- 2025.03: 🎉 Received the UNSW Development and Research Training Grant!
- 2024.07: 🎉 Our survey on AI Safety in Generative AI Large Language Models appeared on ArXiv!
- 2024.07: 🎉 One paper accepted to CIKM 2024!
- 2023.09: 🎉 One paper accepted to ICDM 2023!
- 2022.05: 🎉 Received the UNSW PhD Tuition Fee Scholarship!
- 2022.04: 🎉 Received the CSIRO Data61 PhD Scholarship with a competitive Top-up Award!
- 2022.03: 🎉 Graduated with an MPhil degree from UNSW!
Selected Publications
2026
Causal Abstraction Learning for Multi-Modal Grounded Planning
Xinshu Li, Shiyi Yang, Ziqi Xu, Feng Xia, Michael Sheng, Lina Yao
2026
DrunkAgent: Stealthy Memory Corruption in LLM-Powered Recommender Agents
Shiyi Yang, Zhibo Hu, Xinshu Li, Chen Wang, Tong Yu, Xiwei Xu, Liming Zhu, Lina Yao
2024
Attacking Visually-aware Recommender Systems with Transferable and Imperceptible Adversarial Styles
Shiyi Yang, Chen Wang, Xiwei Xu, Liming Zhu, Lina Yao
2023
Review-Incorporated Model-Agnostic Profile Injection Attacks on Recommender Systems
Shiyi Yang, Chen Wang, Xiwei Xu, Liming Zhu, Lina Yao
2021
Hunter in the Dark: Discover Anomalous Network Activity Using Deep Ensemble Network
Shiyi Yang, Hui Guo, Nour Moustafa
2021
DualNet: Locate Then Detect Effective Payload with Deep Attention Network
Shiyi Yang, Peilun Wu, Hui Guo
2020
Densely Connected Residual Network for Attack Recognition
Peilun Wu, Nour Moustafa, Shiyi Yang, Hui Guo
Selected Awards & Honors
- UNSW Industry Engagement Grant, 2025.
- UNSW Development and Research Training Grant, 2025.
- Outstanding Reviewer, ACM SIGKDD, 2025.
- CSIRO Data61 PhD Scholarship, including competitive Top-up, 2022–2025.
- UNSW PhD Tuition Fee Scholarship (TFS), 2022–2025.
- Provincial-level Outstanding Graduate, 2019.
Invited Talks
- 2026.06 — DrunkAgent: Stealthy Memory Corruption in LLM-Powered Recommender Agents, WWW 2026, Dubai, United Arab Emirates.
- 2024.10 — Attacking Visually-aware Recommender Systems with Transferable and Imperceptible Adversarial Styles, CIKM 2024, Boise, Idaho, USA.
- 2023.12 — Review-Incorporated Model-Agnostic Profile Injection Attacks on Recommender Systems, ICDM 2023, Shanghai, China.
Professional Services
Session Chair: ICDH 2026.
PC Member & Reviewer
- Journals: ACM Transactions on Information Systems (TOIS); IEEE Transactions on Knowledge and Data Engineering (TKDE); IEEE Transactions on Artificial Intelligence (TAI); ACM Transactions on Recommender Systems (TORS); IEEE Transactions on Sensor Networks (TOSN); IEEE Transactions on Big Data (TBD); ACM Transactions on Asian and Low-Resource Language Information Processing (TALLIP); Scientific Reports; etc.
- Conferences: KDD 2024-2027; WWW 2025-2026; NeurIPS 2026; SIGIR 2026; ICDH 2026; ACML 2025; CIKM 2024; etc.
Teaching Experience
Tutor & Mentor at UNSW Sydney since 2020
- COMP9517 Computer Vision, with Prof. Arcot Sowmya, Prof. Erik Meijering, and Dr. Yang Song.
- COMP9444 Neural Networks & Deep Learning, with Dr. Alan Blair.