👋 About Me
I am a fourth-year Ph.D. student in the ML program of Geogia Tech advised by Prof. Pan Li. Prior to that, I received my B.S. degree (Stat, Math and Computer Science) and my M.S. degree (Computer Science) both from Purdue University.
My research focuses on large language models, AI agents, and trustworthy machine learning, with a particular focus on structured information, including graph- and network-structured data, time series, etc.
Specifically, my work spans topics in two main directions:
- LLMs and AI Agents: multi-agent system, KV-cache manipulation, post-training, context management, and agentic data analysis.
- Trustworthy ML: domain adaptation, out-of-distribution generalization, test-time adaptation and robust fine-tuning of foundation models, for reliable graph/geometric learning.
Also, I am currently open to full-time and internship opportunities. Please feel free to reach out about any relevant opportunities!
🔥 News
- 2026.07: 🎉🎉 Our paper Struc-EMB is accepted by COLM’26! Thanks to all my collaborators.
- 2026.06: Our paper Parallel-Synthesis and KVEraser (lead by Mufei) are now available on arXiv. Feel free to check them out!
- 2026.05: I’m starting as a Research Scientist Intern at Meta Central Applied Science team, based in Menlo Park, CA. Would love to connect if you are around:)
- 2026.01: 🎉🎉 Our paper TSA (lead by Hans) gets accepted by AISTATS’26!
- 2025.09: 🎉🎉 Our paper Roft-Mol and the Graph-KV lead by Haoyu gets accepted by NeurIPS’25, thanks to all my collaborators and congratulations to Haoyu and all co-authors! See you in San Diego:)
📝 Publications
Towards Direct Latent-Space Synthesis for Parallel Branches in LLM-Agent Workflows
Shikun Liu, Mufei Li, Dongqi Fu, Haoyu Wang, Yinglong Xia, Hong Li, Hong Yan, Pan Li
KVEraser: Learning to Steer KV Cache for Efficient Localized Context Erasings
Mufei Li, Shikun Liu, Dongqi Fu, Haoyu Wang, Yinglong Xia, Hong Li, Hong Yan, Pan Li
Struc-EMB: The Potential of Structure-Aware Encoding in Language Embeddings
Shikun Liu, Haoyu Wang, Mufei Li, Pan Li.
RoFt-Mol: Benchmarking Robust Fine-Tuning with Molecular Graph Foundation Models
Shikun Liu*, Deyu Zou*, Nima Shoghi, Victor Fung, Kai Liu, Pan Li. NeurIPS 2025 D&B (spotlight)
Graph-KV: Breaking Sequence via Injecting Structural Biases into Large Language Models
Haoyu Wang, Peihao Wang, Mufei Li, Shikun Liu, Siqi Miao, Zhangyang Wang, Pan Li. NeurIPS 2025
Structural Alignment Improves Graph Test-Time Adaptation
Hans Hao-Hsun Hsu*, Shikun Liu*, Han Zhao, Pan Li.
Model generalization on text attribute graphs: Principles with large language models
Haoyu Wang, Shikun Liu, Rongzhe Wei, Pan Li. ICML 2025
Pairwise Alignment Improves Graph Domain Adaptation
Shikun Liu, Deyu Zou, Han Zhao, Pan Li. ICML 2024 (spotlight)
GeSS: Benchmarking Geometric Deep Learning under Scientific Applications with Distribution Shifts
Deyu Zou*, Shikun Liu*, Siqi Miao, Victor Fung, Shiyu Chang, Pan Li. NeurIPS 2024 Dataset and Benchmark
Structural Re-weighting Improves Graph Domain Adaptation
Shikun Liu, Tianchun Li, Yongbin Feng, Nhan Tran, Han Zhao, Qiu Qiang, Pan Li. ICML 2023
Semi-supervised graph neural networks for pileup noise removal
Tianchun Li*, Shikun Liu*, Yongbin Feng*, Garyfallia Paspalaki, Nhan V. Tran, Miaoyuan Liu, Pan Li The European Physical Journal C and NeurIPS 2021 AI4Science
