CAMEO: Curiosity Augmented Metropolis for Exploratory Optimal Policies
This addresses the need for diverse and interpretable optimal policies in reinforcement learning, though it appears incremental as an extension of existing sampling methods.
The paper tackles the problem of multiple optimal policies in reinforcement learning by proposing CAMEO, a curiosity-augmented Metropolis algorithm that samples diverse optimal policies, showing in experiments that it obtains policies solving classic control problems and presenting different risk profiles.
Reinforcement Learning has drawn huge interest as a tool for solving optimal control problems. Solving a given problem (task or environment) involves converging towards an optimal policy. However, there might exist multiple optimal policies that can dramatically differ in their behaviour; for example, some may be faster than the others but at the expense of greater risk. We consider and study a distribution of optimal policies. We design a curiosity-augmented Metropolis algorithm (CAMEO), such that we can sample optimal policies, and such that these policies effectively adopt diverse behaviours, since this implies greater coverage of the different possible optimal policies. In experimental simulations we show that CAMEO indeed obtains policies that all solve classic control problems, and even in the challenging case of environments that provide sparse rewards. We further show that the different policies we sample present different risk profiles, corresponding to interesting practical applications in interpretability, and represents a first step towards learning the distribution of optimal policies itself.