Boosting Robustness in Preference-Based Reinforcement Learning with Dynamic Sparsity

To integrate into human-centered environments, autonomous agents must learn from and adapt to humans in their native settings. Preference-based reinforcement learning (PbRL) can enable this by learning reward functions from human preferences. However, humans live in a world full of diverse information, most of which is irrelevant to completing any particular task. It then becomes essential that agents learn to focus on the subset of task-relevant state features. To that end, this work proposes R2N (Robust-to- Noise), the first PbRL algorithm that leverages principles of dynamic sparse training to learn robust reward models that can focus on task-relevant features. In experiments with a simulated teacher, we demonstrate that R2N can adapt the sparse connectivity of its neural networks to focus on task-relevant features, enabling R2N to significantly outperform several sparse training and PbRL algorithms across simulated robotic environments.

Authors

• Calarina Muslimani
• Bram Grooten Sony AI TU Eindhoven
• Deepak R. S. Mamillapalli University of Alberta
• Mykola Pechenizkiy Eindhoven University of Technology
• Decebal C. Mocanu University of Luxemburg
• Matthew E. Taylor University of Alberta Alberta Machine Intelligence Institute (Amii)

Keywords

• Reinforcement learning
• preference learning
• sparse training