01
Foundation Models
Understanding, training, and evaluating large-scale generative models — from language and vision to multimodal reasoning.
University of Notre Dame · Computer Science & Engineering
Welcome.
Directed by Prof. Xiangliang Zhang, Department of Computer Science and Engineering, University of Notre Dame.
We work on trustworthy foundation models, AI for science, and graph-based learning — with an emphasis on methods that are reliable, interpretable, and broadly applicable.
About the lab
Machine learning research.
Welcome to the MINE lab, directed by Prof. Xiangliang Zhang in the Department of Computer Science and Engineering at the University of Notre Dame.
Our research develops machine learning methods for complex, large-scale data — spanning trustworthy AI, scientific applications in chemistry and biology, and graph-structured problems in drug discovery, recommendation, and knowledge representation.
What we work on
Three research thrusts.
02
AI for Science
Generative and predictive models for chemistry, biology, and physics — accelerating scientific discovery through machine learning.
03
Trustworthy AI
Alignment, safety, fairness, and interpretability of AI systems — building models that are reliable and socially responsible.
Latest
News.
May. 2026
Congratulations to several of our students for their papers accepted by ICML 2026, KDD 2026, ACL 2026, and CHI 2026!
- ProbeLLM: Automating Principled Diagnosis of LLM Failures — ICML 2026
- Agent Training vs Induced Risks! — ICML 2026
- Recipes for Agents: Understanding Skills and Their Open Questions — KDD 2026 Blue Sky Ideas
- Flow Matching + Reaction Prediction — KDD 2026 AI4Sciences
- Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training — ACL 2026
- A Survey of Multimodal Mathematical Reasoning: From Perception, Alignment to Reasoning — ACL 2026
- Guardian-as-an-Advisor: Advancing Next-Generation Guardian Models for Trustworthy LLMs — ACL 2026 Findings
- My Favorite Streamer is an LLM: Discovering, Bonding, and Co-Creating in AI VTuber Fandom — CHI 2026 (Honourable Mention Award)
Mar. 2026
Call for papers! We are co-organizing several workshops: KDD 2026 Workshop on Agentic AI for Scientific and Societal Advances, KDD 2026 Workshop on Reliable Scientific Foundation Models (RelSciFM), AI for Education Day at KDD 2026, CVPR 2026 – 2nd Workshop on Knowledge-Intensive Multimodal Reasoning, and the upcoming ICML 2026 Workshop on AI for Physics (AI4Physics). Welcome submissions!
Jan. 2026
Congratulations to Yue Huang for his papers accepted by ICLR 2026! Check the recent contributions on Benchmarking Generative Foundation Models (TrustGen) and Guardrail for General Agentic Systems. Our work Benchmarking Large Language Models on Safety Issues in Scientific Labs was highlighted by New Scientist and Science News.
Nov. 2025
Congratulations to Yue Huang, Xiaonan Luo, and Xiangqi Wang for their papers accepted by AAAI 2026! Congratulations to Yujun Zhou for the paper Benchmarking Large Language Models on Safety Issues in Scientific Labs, to appear in Nature Machine Intelligence.
Sep. 2025
Congratulations to Yue Huang and Xiangqi Wang for their papers accepted by NeurIPS 2025! We are also organizing an AI for Scientific Research Workshop at AAAI 2026—welcome to submit!
Selected
Recent publications.
ProbeLLM: Automating Principled Diagnosis of LLM Failures
My Favorite Streamer is an LLM: Discovering, Bonding, and Co-Creating in AI VTuber Fandom
TrustGen: A Platform of Dynamic Benchmarking on the Trustworthiness of Generative Foundation Models
Building a Foundational Guardrail for General Agentic Systems via Synthetic Data
Preference Leakage: A Contamination Problem in LLM-as-a-Judge
Benchmarking Large Language Models on Safety Issues in Scientific Labs
Tutorials & workshops
Events we're organizing.
Towards Trustworthy and Socially Responsible Generative Foundation Models
Learn more →Reliable Scientific Foundation Models (RelSciFM)
Learn more →Knowledge-Intensive Multimodal Reasoning (2nd edition)
Learn more →Socially Responsible & Trustworthy Generative Foundation Models
Learn more →Generative Models for Synthetic Data in the GenAI Era
Learn more →Multimodal Mathematical Reasoning
Learn more →
University of Notre Dame · South Bend, Indiana