I am Jun Hou, a PhD student in the Department of Computer Science at Virginia Tech, advised by Dr. Xuan Wang. My research focuses on reliable AI and multimodal learning, with an emphasis on explainability and healthcare applications. My recent work explores how collaborating LLM agents can support clinical prediction and how dependencies among agent errors can inform systematic failure diagnosis. These efforts connect my interests in multimodal modeling, agent systems, and causal reasoning to better understand and improve AI reliability.

Past: Previously, I conducted research on Explainable AI (XAI) in clinical settings under Dr. Lucy Lu Wang , developing techniques to validate and enhance clinicians' trust in NLP-driven clinical models. I hold Master's degrees in Data Analytics Engineering from Northeastern University and Actuarial Science from Columbia University, as well as a Bachelor's in Mathematics from Rutgers University.

Looking ahead, I aim to develop reliable and trustworthy AI systems that account for human factors and the practical challenges of deployment. I am particularly interested in how AI systems can detect and mitigate failures, adapt to evolving tasks and user needs, and support effective human oversight. My goal is to connect technical advances in reliability with how people understand, interact with, and use AI in practice, especially in healthcare and other consequential settings.

I am always open to collaboration and discussions. Please feel free to reach out to explore ideas together or simply chat about research and life!