DoMinO: Discrete Flow-Matching Policy Optimization

Wed, 01 Apr 2026 00:00:00 +0000

Role: Project Co-Lead · Northwestern University (Remote) · Feb 2026 – Apr 2026

DoMinO is a reinforcement learning framework for fine-tuning discrete flow-matching models toward a reward.

What I did

Co-developed the framework. DoMinO formulates discrete flow-matching inference as an inner Markov decision process, which enables exact trajectory log-likelihood optimization.
Implemented the fine-tuning methods. REINFORCE- and PPO-based fine-tuning with cross-entropy and generalized KL regularization, to mitigate reward hacking and preserve sequence naturalness.
Evaluated on biological sequence design. On regulatory DNA and protein design, DoMinO outperformed prior state-of-the-art baselines by 10.6% on DNA predicted activity and 15.4% on protein pred-ddG.

Paper: . See also the .