Research

LLM Alignment

• Mix-and MoE-DPO: A Variational Inference Approach to Direct Preference Optimization.
  Jason Bohne, Pawel Polak, David Rosenberg, Brian Bloniarz, and Gary Kazantsev. Preprint, 2025.

• Learning to Trade with Preferences: Interpretable Execution via Mixture-of-Experts.
  Haohan Xu, Jason Bohne, Pawel Polak, Dave Byrd, David Rosenberg, and Gary Kazantsev. ICAIF 2025.

Bilevel Optimization

• Non-Stationary Functional Bilevel Optimization.
  Jason Bohne, Ieva Petrulionyte, Michael Arbel, Julien Mairal, and Pawel Polak. AISTATS 2026.

• Bregman Geometry for Stochastic Online Bilevel Optimization.
  Jason Bohne, David Rosenberg, Gary Kazantsev, and Pawel Polak. Preprint, 2025.

• Online Nonconvex Bilevel Optimization with Bregman Divergences.
  Jason Bohne, David Rosenberg, Gary Kazantsev, and Pawel Polak. NeurIPS OPT Workshop, 2024.

Financial AI

• When Quotes Crumble: Detecting Transient Mechanical Liquidity Erosion in Limit Order Books.
  Haohan Xu, Jason Bohne, Pawel Polak, Yuri Baransky, Alberto Alva, Valentina Fedotova, and Gary Kazantsev. 2nd Workshop on Advances in Financial AI (ICLR 2026).
  🏆 Best Poster Award, Runner-Up (ICLR 2026).

For a full list, see my Google Scholar profile.