OPTIMIZATION · LEARNING · STATISTICS · OPERATIONS RESEARCH

Yifan Hu 逸凡

Assistant Professor

Department of Statistics · Rutgers University
New Brunswick, New Jersey

I study decision making under uncertainty through optimization, learning, operations research, and statistics.

Portrait of Yifan Hu
YIFAN HURUTGERS UNIVERSITY · 2026
FocusReliable and scalable learning and decision making
Methods
  • Optimization
  • Reinforcement Learning
  • Generative AI
Applications
  • Causal Discovery
  • Scientific Discovery
  • Supply Chain
  • Learning Systems
  • Game Design
OpportunitiesPhD positionsResearch positions

01 / RESEARCH

Research Directions.

Foundations and practical methods for reliable decisions in complex systems.

01

Stochastic Optimization

Global solutions through hidden convexity, and efficient methods for bilevel and multistage problems with biased gradients.

Explore work
02

Reinforcement Learning

Robust and contextual decision making, including policy optimization and bilevel learning.

Explore work
03

Causality & Statistics

Optimization for causal discovery and instrumental-variable inference, alongside robust learning and fairness.

Explore work
04

Learning Systems

Test-time computation, robust preference optimization, and on-policy distillation.

Explore work

02 / UPDATES

Latest News.

Research, talks, and milestones from the group.

  1. Two new preprints are out: PRICE studies how to allocate test-time rollouts and voting under a token budget, while Dr. OPD studies which teacher signals improve on-policy distillation.

  2. Our PRICE paper was selected for an oral presentation at the NeurIPS 2026 FLLMPT Workshop and accepted to its MLxOR, TTCL, and RAAAI workshops.

  3. A new preprint with Mengmeng Li, Daniel Kuhn, and Yan Li develops sample-efficient methods for robust Markov decision processes with continuous state spaces.

  4. A new preprint with Buse Şen and Daniel Kuhn introduces multistage conditional compositional optimization and reduces the scenario cost of nested decisions.

  5. I received the 2025 Mathematical Programming Meritorious Service Award.

03 / WRITING

Blog.

All posts ↗

JOIN THE GROUP

Working With Us.

I welcome motivated PhD applicants, research interns, thesis students, and collaborators interested in optimization, learning, and decision making.

Explore Opportunities