Zhan Lab

Welcome to the Zhan Lab at New Mexico Tech!

We build machine learning systems that connect large language models and foundational models with real-world scientific and societal challenges. Our work spans deep learning, structured reasoning, and human-AI interaction, with applications across genomics, medicine, robotics, and responsible AI. We are especially interested in agentic AI and foundation models that advance scientific discovery and biomedicine.

New Mexico Tech – Institute Flag

🔬 Research Areas

• 🧬 Variant Effect Prediction using Large Language Models
  We design LLMs to interpret genetic variants and predict disease relevance across cardiomyopathies and neurological disorders.

• 🧠 Biomedical Knowledge Graphs & Retrieval-Augmented Reasoning
  Hybrid methods that combine neural models with structured biological knowledge.

• 🔐 Privacy & Robustness in Genomic Models
  Studying adversarial attacks, fairness, and model bias for real-world deployment.

• 🤖 Agentic AI for Surgery and Rehab
  AI systems that analyze and collaborate in clinical actions such as surgical skill tracking and robotic rehabilitation.

• 🔬 Agentic AI & Foundation Models for Scientific Discovery and Biomedicine
  LLM-based agents and foundation models that plan, reason, and use tools to accelerate hypothesis generation and biomedical discovery.

Recent News

• [May, 2026]. Awarded NM Research Infrastructure Development (RID) from NASA / New Mexico Space Grant Consortium as PI for Disentangling Radiation- and Microgravity-Driven Gene Expression Programs in Public Space Biology Datasets (2026–2027)!
• [Jan, 2026]. Awarded NM-INBRE Developmental Research Project Program (DRPP) as PI for REACT: Rehabilitation through Embodied Agentic Clinical Transfer (2025–2026)!
• [Dec, 2025]. Awarded Supporting Undergraduate Research Experiences (SURE) from New Mexico EPSCoR as PI for Large Language Models for Code and Compiler Education: A Course-Based Undergraduate Research Experience (2025–2026)!
• [Mar, 2025]. "A disease-specific language model for variant pathogenicity in cardiac and regulatory genomics" is published in Nature Machine Intelligence!
• [Mar, 2025]. Invited to serve as a reviewer for ACL-2025 and Nature Scientific Reports!
• [Jun, 2024]. Invited to serve as a PC member for AAAI-2025 and a reviewer for both Machine Learning and Autonomous Agents and Multi-Agent Systems!
• [Jan, 2024]. Invited to serve as a TPC member for IJCNN-2024, and a reviewer for ACM Transactions on Human-Robot Interaction!
• [Oct, 2023]. Invited to serve as a PC member for AAAI-2024, SDM-2024, and ICKG-2023!
• [Aug, 2023]. "Simplex2vec Backward: From Vectors Back to Simplicial Complex" is accepted by CIKM-2023!
• [Jul, 2023]. Invited to serve as a PC member for IEEE-SMC 2023 and HICSS-2024, as well as a subreviewer for CIKM-2023!
• [Jul, 2023]. Start a new job as a Postdoctoral Scientist at Cedars-Sinai!
• [Apr, 2023]. Invited to serve as a PC member for IJCAI-2023!
• [Dec, 2022]. Received the AAAI-23 Student Scholarship!
• [Dec, 2022]. Invited to serve as a Program Committee (PC) member for KDD-2023 Research Track!
• [Dec, 2022]. "Measuring the Privacy Leakage via Graph Reconstruction Attacks on Simplicial Neural Networks" has been selected as one of the finalists!
• [Nov, 2022]. "Measuring the Privacy Leakage via Graph Reconstruction Attacks on Simplicial Neural Networks" is accepted by AAAI-2023 Student Abstract and Poster Program!
• [Nov, 2022]. "Privacy-Preserving Representation Learning for Text-Attributed Networks with Simplicial Complexes" is accepted by AAAI-2023 Doctoral Consortium!
• [Nov, 2022]. "Towards Fair and Selectively Privacy-Preserving Models Using Negative Multi-Task Learning" is accepted by AAAI-2023 Student Abstract and Poster Program!

📚 Featured Publications

A disease-specific language model for variant pathogenicity in cardiac and regulatory genomics Huixin Zhan, Jason H. Moore, Zijun Frank Zhang Nature Machine Intelligence, 2025