Xilin Dang

I am a Ph.D. candidate at Computer Science and Engineering from The Chinese University of Hong Kong, supervised by Prof. HENG Pheng-Ann. Previously, I received my B.Eng. degree in Computer Science and Technology in Jinan University, supervised by Prof. Ke Wang and Prof. Zhen Zhang.

My research interests cover Agentic AI and AI applications in biomedicine.

Email  /  Codeforces  /  Google Scholar  /  Github

profile photo
Research

I'm interested in agentic AI. I want to try doing work that is both meaningful and interesting.

Advancing Radiograph Representation Learning via Cascading Graph Alignment for Vision-Language Clinical Concepts
Xilin Dang, Kang Li, Pheng Ann Heng
Medical Image Analysis, 2026

An vision-language alignment framework to enable explicit alignment for the clinically significant concepts as well as the diagnostic content summary.

Knowguard: Knowledge-driven abstention for multi-round clinical reasoning
Xilin Dang*, Kexin Chen*, Xiaorui Su, Ayush Noori, Inaki Arango, Lucas Vittor, Xinyi Long, Yuyang Du, Marinka Zitnik, Pheng Ann Heng
ICLR, 2026

An interactive LLM agent using multimodal knowledge graphs for Uncertainty-Aware Question Answering and Multi-Hop Reasoning with proactive question asking capabilities under incomplete or ambiguous inputs.

Uncovering Hidden Vulnerabilities in Convolutional Neural Networks through Graph-based Adversarial Robustness Evaluation
Ke Wang, Zicong Chen, Xilin Dang, Xuan Fan, Xuming Han, Chien-Ming, WeiPing Ding, Siu-Ming Yiu, Ke Wang
Pattern Recognition, 2023

This paper proposes a new robustness quantitative measurement scheme from a macroscopic perspective, modeling CNN as graph and using complex system theory.

A Defense Method based on Attention Mechanism against Traffic Sign Adversarial Samples
Hailiang Li, Bin Zhang, Yu Zhang, Xilin Dang, Yuwei Han, Linfeng Wei, Yijun Mao
INFORMATION FUSION, 2020

Inspired by the principle of human vision, this paper proposes a defense method based on an attentional mechanism for traffic sign adversarial samples.