CV

Current academic CV

Publications, research, education, and academic service.

Download PDF

Research Interests

Reinforcement learning and generative models; optimization and control for decision-making under uncertainty.

Education

Fudan University, Shanghai, China

B.S. in Information and Computing Science & Artificial Intelligence (Double Degree), School of Mathematical Sciences

GPA 3.98/4.00 · Rank 3/228

Publications

Calibrating What Behavior Can Identify: Conformal Reward Sets for Inverse Reinforcement Learning
Yuhan Chi. NeurIPS 2026 Workshop on Physical Understanding for Decision-Making (PUDM), Poster.
Paper · Code · Project

  • Adapts conformal inverse optimization to IRL with a coordinate-invariant, policy-value-based score for calibrated reward sets.
  • Characterizes when behavioral compatibility implies true-reward coverage, and proves when robust planning over the set cannot certify improvement over a reference policy.

Token-Level Verification under Controlled Evaluation
Yuhan Chi. ICML 2026 Workshop on AI for Math (AI4Math), Poster.
Paper · Code

  • Shows that evaluation-protocol choices alone shift apparent verification AUROC of LLM math traces by up to 0.18.
  • Under controlled scoring, shallow token statistics reach only 0.60–0.75 AUROC; final-token entropy drops from 0.72–0.75 to chance (0.47–0.48).

Research Experience

Visiting Student, Mechanical Systems Control Lab, UC Berkeley

Studied robot hardware interfaces and explored dexterous-hand grasping in simulation.

Selected Projects

code-not-text — cross-domain measurement study, 31,040 reasoning runs · Demo

  • Selecting among 64 candidates with cheap reasoning-trace features changes pass@1 by +10.0 pp (math), +8.0 pp (science), and −0.6 pp (code); rules out five alternative explanations for the coding failure.

TinyLoRA-GRPO-Coder — RL post-training for code generation

  • Trains 32 shared scalars of Qwen2.5-Coder-3B with GRPO, rewarded by real g++ compile-and-run outcomes.

Honors, Talks & Service

Skills

Programming: Python, C++, C
ML: PyTorch, Machine Learning, Reinforcement Learning, Generative Models
Tools: Git, Linux, LaTeX
Languages: Chinese (native), English (fluent)

Selected Coursework

Mathematical Analysis I, II (A, A+) · Advanced Algebra I, II (A−, A) · Analytic Geometry (A) · Programming (A) · Introduction to AI (A) · Foundations of Software for AI (A)

Open Source & Community

Founder, Fudan Open Source Initiative
Projects: FDUGuideBook, FDU-Sharing, ic-guide, github-unflag-playbook-cn