Hi There, Welcome!
I'm an incoming assistant professor at the Hong Kong University of Science and Technology (HKUST) Computer Science & Engineering (CSE) Department, starting in Fall'24. I am actively recruiting PhD, Masters, and Undergraduate student to join us in exciting NLP/AI research work. If you're interested, please fill out this form, thanks!
At present, I am a Ph.D. candidate at the Blender Lab of UIUC, where I am fortunately advised by Prof. Heng Ji. My research focus centers around socially-situated human-centered trustworthy NLP/AI with multimedia knowledge reasoning capability and scalable alignment principles. In particular, here’s a handful selection of my recent research that dive into the fundamental model architectural mechanisms or frameworks empowering harmless, helpful, and honest information communication:
- Norms Discovery and Explainable Norms Violation Detection: NormSAGE[1]; LM debiasing and detoxification[2][3]
- Fine-Grained Information Extraction: NewsClaim, a novel benchmark for claim extraction with attribute knowledge[4]; RESIN, a state-of-the-art schema-guided cross-document cross-media IE and event tracking system[5]
- Misinformation Detection: reasoning across multimedia and background knowledge[6]; cross-document misinformation detection [7]; battling fake and biased news[8]
- Diverse Human-Centered Pro-Social Reasoning Tasks: SmartBook generation for situation reports[9]; Decoding the Silent Majority, for social media response forecasting[10]; COVID KG Construction and Drug Report Generation[11]
- Intelligent Agent with Tool Learning Capability: CREATOR, LLM tool creation through abstraction reasoning[12]; CRAFT, customizing LLMs through specialized toolset creation and retrieval[13]
- Foundation Model Hallucination Control and Alignment in Honesty/Faithfulness: R-Tuning, teaching LLMs to refuse uncertain questions[14]