Mo and Weerasooriya publish on Stability-Augmented Reinforcement Policy Optimization

Fengran Mo, assistant professor in the Department of Computer Science, and Cyril Weerasooriya '24 (computer science) co-authored “StaRPO: Stability-Augmented Reinforcement Policy Optimization” with peers from other universities. The paper, which was accepted to this year's Conference on Neural Information Processing Systems, explains StaRPO, a stability-augmented reinforcement learning framework that explicitly incorporates reasoning stability into the optimization objective.

Related News