Sizhe Liu

Hello! I’m a senior undergraduate at USC studying computer science. I am a student researcher at Liang’s Lab, advised by Prof. Liang Chen, and at FORTIS Lab, advised by Prof. Yue Zhao. I also work closely with Prof. Tianfan Fu and Prof. Tea Jashashvili.
Additionally, I’m a visiting student at Westlake University, working with Jun Xia under the supervision of Prof. Stan Z. Li.
My research focuses on developing machine learning methods for biomedical applications, with key areas including:
- ML models for biomedical discovery: Developing geometric and generative AI models for biological tasks such as de novo peptide sequencing, gene perturbation, drug-target interaction, etc.
- ML systems: Designing and building ML systems that make these discoveries efficient, reliable, and trustworthy.
- Automating scientific research: Using LLM agents to streamline and automate scientific research workflows.
news
Jan 23, 2025 | One paper have been accpeted by ICLR 2025 ![]() |
---|---|
Jan 11, 2025 | We have a new paper published in Bioinformatics ![]() |
Dec 20, 2024 | We have a new preprint on LLM agents for drug discovery |
Oct 15, 2024 | One paper accepted at AIDrugX@NeurIPS 2024 (Spotlight) |
Sep 26, 2024 | Two papers have been accpeted by NeurIPS 2024 ![]() |
selected publications
- HMGDeciphering single-cell gene expression variability and its role in drug responseHuman Molecular Genetics, Sep 2024
- NeurIPSFlexMol: A Flexible Toolkit for Benchmarking Molecular Relational LearningIn Neural Information Processing Systems (NeurIPS) Track on Datasets and Benchmarks, Sep 2024
- BioinformaticsSP-DTI: Subpocket-Informed Transformer for Drug-Target Interaction PredictionBioinformatics, Jun 2025
- AI4ResearchDrugAgent: Automating AI-aided Drug Discovery Programming through LLM Multi-Agent CollaborationIn AI4Research@AAAI, Dec 2024
- ICLRBridging the Gap between Database Search and De Novo Peptide Sequencing with SearchNovoInternational Conference on Learning Representations, Oct 2025
- AABAMulti-Bone Micro-CT Scanning with a Leap Towards Reproducibility and Efficiency Using Automated ToolsIn American Association of Biological Anthropologists (AABA) Conference, Mar 2024
- NeurIPSNovoBench: Benchmarking Deep Learning-based \emphDe Novo Sequencing Methods in ProteomicsIn Neural Information Processing Systems (NeurIPS) Track on Datasets and Benchmarks, Sep 2024