Active Reinforcement Learning Strategies for Offline Policy Improvement
This addresses the challenge of limited interaction budgets in RL for applications like medical trials and navigation, though it appears incremental as it builds on existing offline RL and active learning concepts.
The paper tackles the problem of expensive online interactions in reinforcement learning by proposing an active reinforcement learning method that collects minimal additional trajectories to augment offline data, reducing additional online interaction by up to 75% across various environments.
Learning agents that excel at sequential decision-making tasks must continuously resolve the problem of exploration and exploitation for optimal learning. However, such interactions with the environment online might be prohibitively expensive and may involve some constraints, such as a limited budget for agent-environment interactions and restricted exploration in certain regions of the state space. Examples include selecting candidates for medical trials and training agents in complex navigation environments. This problem necessitates the study of active reinforcement learning strategies that collect minimal additional experience trajectories by reusing existing offline data previously collected by some unknown behavior policy. In this work, we propose an active reinforcement learning method capable of collecting trajectories that can augment existing offline data. With extensive experimentation, we demonstrate that our proposed method reduces additional online interaction with the environment by up to 75% over competitive baselines across various continuous control environments such as Gym-MuJoCo locomotion environments as well as Maze2d, AntMaze, CARLA and IsaacSimGo1. To the best of our knowledge, this is the first work that addresses the active learning problem in the context of sequential decision-making and reinforcement learning.