Deep Autoregressive Density Nets Vs Neural Ensembles For Model-based Offline Reinforcement Learning

Abstract

We consider the problem of offline reinforcement learning where only a set of system transitions is made available for policy optimization. Following recent advances in the field, we consider a model-based reinforcement learning algorithm that infers the system dynamics from the available data and performs policy optimization on imaginary model rollouts. This approach is vulnerable to exploiting model errors which can lead to catastrophic failures on the real system. The standard solution is to rely on ensembles for uncertainty heuristics and to avoid exploiting the model where it is too uncertain. We challenge the popular belief that we must resort to ensembles by showing that better performance can be obtained with a single well-calibrated autoregressive model on the D4RL benchmark. We also analyze static metrics of model-learning and conclude on the important model properties for the final performance of the agent.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keybenechehab2024deep

Related papers

• Revisiting Design Choices In Offline Model-based Reinforcement Learning (2021)6.34
• Overcoming Model Bias For Robust Offline Deep Reinforcement Learning (2020)11.58
• Learning From Sparse Offline Datasets Via Conservative Density Estimation (2024)0.00
• Balancing Policy Constraint And Ensemble Size In Uncertainty-based Offline Reinforcement Learning (2023)5.24
• D4RL: Datasets For Deep Data-driven Reinforcement Learning (2020)0.00
• An Optimistic Perspective On Offline Reinforcement Learning (2019)0.00
• Deployment-efficient Reinforcement Learning Via Model-based Offline Optimization (2020)0.00
• Regularizing A Model-based Policy Stationary Distribution To Stabilize Offline Reinforcement Learning (2022)0.00