Scholar
Han Shen
Google Scholar ID: UeWSr6oAAAAJ
Research Engineer, Ant Group; Ph.D., Rensselaer Polytechnic Institute
Citations & Impact
Citations
463
H-index
9
i10-index
9
Publications
17
Co-authors
13
list available
Publications
Resume (English only)
Academic Achievements
- Paper 'AEnt' published, its asynchronous implementation is incorporated in the highly scalable RL framework AReaL.
- Paper 'SEAL: Safety-enhanced Aligned LLM Fine-tuning via Bilevel Data Selection' accepted at ICLR 2025.
- Paper 'Principled Penalty-based Methods for Bilevel Reinforcement Learning and RLHF' accepted at ICML 2024, extended work in JMLR.
- Paper 'Mitigating Forgetting in LLM Supervised Fine-Tuning and Preference Learning' published.
- Paper 'On Penalty-based Bilevel Gradient Descent Method' accepted at ICML 2023, extended work in Mathematical Programming.
- Paper 'Mitigating Gradient Bias in Multi-objective Learning: A Provably Convergent Approach' accepted as an oral presentation at ICLR 2023.
Education
- Ph.D. from RPI, supervised by Dr. Tianyi Chen (now at Cornell Tech). He was the first Ph.D. student in Dr. Tianyi Chen's group, focusing on optimization and reinforcement learning.
Background
- Currently a senior research engineer at Ant Group, working on a variety of LLM alignment and reinforcement learning. Previously, he worked as a research intern at IBM Research AI, collaborating with Pin-Yu Chen, Payel Das, Songtao Lu, Xiaodong Cui, and many other talented researchers. His research at IBM focused on LLM alignment and offline RL.
Miscellany
- Reviewer for NeurIPS, ICML, ICLR, AISTATS, AAAI, and IEEE Transactions on Signal Processing (TSP).
