4 papers · 1 filter
In-Distribution Imagination for Model-Based Offline Reinforcement Learning
Mintae Kim, Koushil Sreenath
Model-based offline reinforcement learning (MBORL) improves sample efficiency through model-generated trajectories. However, accumulative model error can drive imagined trajectorie…
Trajectory-Regularized Stochastic Optimal Control via KL Divergence
Mintae Kim, Koushil Sreenath
We introduce trajectory-regularized stochastic optimal control (TRSOC), which augments standard stochastic optimal control (SOC) with a Kullback--Leibler (KL) divergence between co…
Robust Adversarial Policy Optimization Under Dynamics Uncertainty
Mintae Kim, Koushil Sreenath
Reinforcement learning (RL) policies often fail under dynamics that differ from training, a gap not fully addressed by domain randomization or existing adversarial RL methods. Dist…
WOMBET: World Model-Based Experience Transfer for Robust and Sample-efficient Reinforcement Learning
Mintae Kim, Koushil Sreenath
Reinforcement learning (RL) in robotics is often limited by the cost and risk of data collection, motivating experience transfer from a source task to a target task. Offline-to-onl…