DIAR: Diffusion-model-guided Implicit Q-learning with Adaptive Revaluation
This addresses offline RL problems for researchers and practitioners, offering incremental improvements in handling long-horizon, sparse-reward environments.
The paper tackles challenges in offline reinforcement learning, such as out-of-distribution samples and long-horizon problems, by proposing the DIAR framework, which outperforms state-of-the-art algorithms in tasks like Maze2D, AntMaze, and Kitchen.
We propose a novel offline reinforcement learning (offline RL) approach, introducing the Diffusion-model-guided Implicit Q-learning with Adaptive Revaluation (DIAR) framework. We address two key challenges in offline RL: out-of-distribution samples and long-horizon problems. We leverage diffusion models to learn state-action sequence distributions and incorporate value functions for more balanced and adaptive decision-making. DIAR introduces an Adaptive Revaluation mechanism that dynamically adjusts decision lengths by comparing current and future state values, enabling flexible long-term decision-making. Furthermore, we address Q-value overestimation by combining Q-network learning with a value function guided by a diffusion model. The diffusion model generates diverse latent trajectories, enhancing policy robustness and generalization. As demonstrated in tasks like Maze2D, AntMaze, and Kitchen, DIAR consistently outperforms state-of-the-art algorithms in long-horizon, sparse-reward environments.