Evolutionary Bilevel Reward Shaping for Generalization in Reinforcement Learning
This paper proposes a bilevel optimization approach to improve generalization of reinforcement learning in unseen test environments using only scalar feedback from validation environments.
The paper presents a novel bilevel optimization approach that improves generalization in reinforcement learning without requiring access to diverse training environments or full trajectory observability.
Before reading this…
Applications
- →Improving generalization in reinforcement learning
To understand this paper, make sure you know these concepts first:
- Reinforcement learningfind papers →
- Bilevel optimizationfind papers →
Abstract
More Like ThisReinforcement learning (RL) often suffers from performance degradation when deployed in environments that differ from those encountered during training. Existing techniques such as domain randomization (DR) mitigate this, but require access to diverse training environments and full trajectory observability, assumptions that fail in privacy-preserving or restricted scenarios where only scalar performance metrics are available. We propose Generalization via Evolutionary Reward Shaping (GERS), a bilevel optimization approach to improve generalization on unseen test environments using only scalar feedback from validation environments. At the lower level, an RL agent guided via a reward function shaped by the upper level learns a policy on a limited set of training environments with accessible trajectory data; at the upper level, CMA-ES optimizes the reward shaping parameters to maximize the cumulative unshaped reward on separate validation environments for which trajectory access is unavailable. Results on continuous control tasks indicate that GERS outperforms the standard RL baseline on unseen test environments. GERS performance is comparable to DR, despite DR treating the combined set of training and validation environments of GERS as a single training set that requires trajectory access, whereas GERS cannot access validation trajectories. These results confirm that GERS effectively enhances generalization under restricted data access constraints.