Xuefeng Liu

Incoming tenure-track Assistant Professor, University of Florida
Postdoc, School of Medicine, Stanford University
Ph.D., Department of Computer Science, University of Chicago
xuefeng.liu@ufl.edu (primary), xfl@stanford.edu, xuefeng@uchicago.edu

prof_pic2.jpg

265 Campus Drive

Stanford, CA 94305

Research: My research spans two complementary directions.

  • Machine Learning Foundations: I develop practically driven, theoretically grounded methods in reinforcement learning, generative modeling and Agentic AI.

  • AI for Biomedicine: I integrate biomedical modeling with advanced AI to address the CURED challenges — Controllability, Unified multimodality, Robustness, Efficiency, and Dependability on biophysical and biological principles — in disease diagnosis and therapy.

My research interests include, but are not limited to:

  • AI for Biomedicine
    • Biomolecular design, drug discovery, biomarker discovery, lead discovery and optimization
  • Reinforcement Learning
    • RL in pretraining, post-training, decoding
    • Reasoning, planning, and decision-making under uncertainty
  • Agentic AI for Scientific Discovery
    • Multimodal reasoning systems for autonomous scientific discovery
  • Generative AI and Foundation Models
    • Generative modeling for biomedicine and beyond
  • Scientific Machine Learning
    • Physics- and Biology-informed machine learning
  • Human-AI co-scientist

About Me: I am a Postdoctoral Fellow at School of Medicine, Stanford University, working with Prof. Le Cong and Prof. Mengdi Wang (Princeton University). Before joining Stanford, I received my Ph.D. in Computer Science from University of Chicago, where I was advised by Prof. Rick L. Stevens, with co-advisors Prof. Yuxin Chen and Prof. Jinbo Xu, and mentorship from Prof. Tobin R. Sosnick. I also served as a Research Associate at Argonne National Laboratory, where my work focuses on AI for Biomedicine.

Open Opportunities:

  • [Recruiting] Starting in Fall 2026, I will join the University of Florida as a tenure-track Assistant Professor, and I am actively recruiting PhD students and Postdoc researchers. I am looking for highly motivated candidates with a passion for research and strong coding skills; solid mathematical foundations are a plus. Please send your CV, a brief summary of your research experience, and a description of your research interests to xuefeng.liu@ufl.edu.

  • [Research Students] I have open research opportunities—feel free to email me if you are interested! Candidates are expected to have earned an A- or A in Deep Learning, or a related course.

Teaching:

  • [CAI 6734] Applied Generative AI in Medicine — Fall 2026

news

Jul 04, 2026
  • Our paper "DrugImproverGPT: GPT-Driven Drug Optimization with Structured Policy Optimization Post-training" is accepted by Machine Learning for Healthcare (MLHC) 2026! Thanks to my collaborators Songhao, Siyu, Zhuoran, Yuxin, Ian, and Rick!
Jul 03, 2026
Jul 01, 2026
  • Our preprint "Active-GRPO: Adaptive Imitation and Self-Improving Reasoning for Molecular Optimization" is posted on arXiv.
Jun 20, 2026
  • Our paper "Regime-Adaptive Bayesian Optimization via Dirichlet Process Mixtures of Gaussian Processes" is accepted by ICML 2026.
May 31, 2026
  • Our preprint "Protein Thoughts: Interpretable Reasoning with Tree of Thoughts and Embedding-Space Flow Matching for Protein-Protein Interaction Discovery" is posted on arXiv.
May 15, 2026
  • My Ph.D. thesis "Leveraging Active Reinforcement Learning and Generative Models for Biomolecular Design" is completed at the University of Chicago.
Apr 15, 2026
  • Our paper "CACHE Challenge# 3: Targeting the Nsp3 Macrodomain of SARS-CoV-2" is published in Journal of Chemical Information and Modeling.
Feb 20, 2026
  • Our paper "Multi-Objective Coverage via Constraint Active Search" is accepted by AAMAS 2026 as an oral presentation.
Dec 02, 2025
  • Traveling to San Diego to attend NeurIPS 2025.
    • I am organizing the workshop AI Virtual Cells and Instruments: A New Era in Drug Discovery and Development. In light of the FDA’s recent initiative to phase out animal testing requirements, the workshop aims to foster deeper discussion on AI virtual cells and computational instruments for drug discovery. Looking forward to seeing you on Saturday, December 6.
    • Organizers: Quanquan Gu (UCLA), Michelle M. Li (Harvard), Chong Liu (UAlbany), Xuefeng Liu, Abhishek Pandey (AbbVie), Ji Won Park (Prescient Design, Genentech), Natasa Tagasovska (Prescient Design, Genentech), and Marinka Zitnik (Harvard).
    • Invited speakers: Linda Goodman (FaunaBio), Arvind Ramanathan (Argonne National Laboratory), Mengdi Wang (Princeton), Eric Xing (MBZUAI, GenBio, & CMU), Jinbo Xu (TTIC & Molecule Mind), and Alex Zhavoronkov (Insilico Medicine).
Nov 20, 2025
  • Our paper "FragmentGPT: A Unified GPT Model for Fragment Growing, Linking, and Merging in Molecular Design" is selected as an oral presentation at the NeurIPS 2025 Workshop on AI Virtual Cells and Instruments.
Nov 10, 2025
  • Our paper "Monte Carlo Tree Diffusion with Multiple Experts for Protein Design" is presented at the NeurIPS 2025 Workshop on AI Virtual Cells and Instruments.
Aug 15, 2025
  • Our paper "Bidirectional Hierarchical Protein Multi-Modal Representation Learning" is accepted by MLHC 2025.
Aug 10, 2025
  • Our paper "ScaffoldGPT: A Scaffold-based GPT Model for Drug Optimization" is accepted by MLHC 2025.
Jul 20, 2025
  • Our paper "A Ground-Up Designed Controllable GPT for Molecule Optimization" is presented at the ICML 2025 Generative AI and Biology Workshop.
Jul 15, 2025
  • Our paper "Scaffold-Driven GPT Model for Drug Optimization" is presented at the ICML 2025 Generative AI and Biology Workshop.
Jul 10, 2025
  • Our paper "Active Advantage-Aligned Online Reinforcement Learning with Offline Data" is presented at the ICML 2025 Exploration in AI Today Workshop.
Apr 04, 2025
Feb 15, 2025
  • Our preprint "ControllableGPT: A Ground-Up Designed Controllable GPT for Molecule Optimization" is posted on arXiv.
Nov 25, 2024
  • Our preprint "Reflections from the 2024 Large Language Model (LLM) Hackathon for Applications in Materials Science and Chemistry" is posted on arXiv.
Nov 15, 2024
  • Our paper "Contextual Active Model Selection" is accepted by NeurIPS 2024.
Oct 10, 2024
  • Our preprint "Binding Affinity Prediction: From Conventional to Machine Learning-Based Approaches" is posted on arXiv.
May 15, 2024
  • Our paper "Entropy-Reinforced Planning with Large Language Models for Drug Discovery" is accepted by ICML 2024.
May 05, 2024
  • Our paper "Learning from Imperfect Human Feedback: A Tale from Corruption-Robust Dueling" is accepted by ICLR 2024 and received the Best Poster Award at Midwest ML Symposium 2024.
Apr 20, 2024
  • Our paper "Leveraging Protein Large Language Models and Graph Neural Networks for Binding Affinity Prediction" is presented at the ICML 2024 Machine Learning for Life and Material Science Workshop.
Apr 10, 2024
  • Our paper "APO: Advantage-Alignment Policy Optimization for Fine-Tuning Generative Models" is presented at the ICLR 2024 Generative Models for Decision Making Workshop.
Dec 05, 2023
  • Our paper "DRUGIMPROVER: Utilizing Reinforcement Learning for Multi-Objective Alignment in Drug Optimization" is selected as an oral presentation at the NeurIPS 2023 Workshop on New Frontiers of AI for Drug Discovery and Development.
Jul 20, 2023
  • Our paper "Active Policy Improvement from Multiple Black-Box Oracles" is accepted by ICML 2023.
May 10, 2023
  • Our paper "Blending Imitation and Reinforcement Learning for Robust Policy Improvement" is accepted by ICLR 2023 as a spotlight presentation.
Feb 10, 2023
  • Our paper "AI-Accelerated Protein-Ligand Docking for SARS-CoV-2 Is 100-Fold Faster with No Significant Change in Detection" is published in Scientific Reports.