
Towards Robust Multimodal Large Language Models Against Jailbreak Attacks
Ph.D. Student (4th year) · PSU Data Science Lab
The Penn State University, PA, United States
Email: ziyiyin1998 [at] gmail [dot] com
I am a fourth-year Ph.D. student supervised by Prof. Fenglong Ma and Prof. Ting Wang at the PSU Data Science Lab. Currently, I'm working on multimodal learning and AI in security. I received my M.S. and B.S. degrees from the Department of Automation Science and Engineering (Youth Class) at Xi'an Jiaotong University (XJTU) in 2022 and 2019.

Towards Robust Multimodal Large Language Models Against Jailbreak Attacks

Shadow-Activated Backdoor Attacks on Multimodal Large Language Models

VQAttack: Transferable Adversarial Attacks on Visual Question Answering via Pre-trained Models

VLAttack: Multimodal Adversarial Attacks on Vision-Language Tasks via Pretrained Models
ICDAGENT: Empowering Agentic Large Language Models for Explainable Medical Coding

Asymmetrical Reciprocity-based Federated Learning for Resolving Disparities in Medical Diagnosis
*Equal contribution

Multimodal Transformer Network for Pedestrian Trajectory Prediction

Order-independent Matching with Shape Similarity for Parking Slot Detection

Learning to Plan Semantic Free-Space Boundary
Applied Scientist Intern — Active Interaction and Perception Reasoning in VLA Models
Mentors: Sangmin Woo & Kang Zhou
Applied Scientist Intern — Towards Parameter-Efficiency on SMoE LLMs
Mentors: Zhengyang Wang & Hejie Cui