Adaptive Curriculum Learning With Successor Features for Imbalanced Compositional Reward Functions

Laszlo Szoke, Shahaf S. Shperberg, Jarrett Holtz, Alessandro Allievi

Research output: Contribution to journalArticlepeer-review

Abstract

This work addresses the challenge of reinforcement learning with reward functions that feature highly imbalanced components in terms of importance and scale. Reinforcement learning algorithms generally struggle to handle such imbalanced reward functions effectively. Consequently, they often converge to suboptimal policies that favor only the dominant reward component. For example, agents might adopt passive strategies, avoiding any action to evade potentially unsafe outcomes entirely. To mitigate the adverse effects of imbalanced reward functions, we introduce a curriculum learning approach based on the successor features representation. This novel approach enables our learning system to acquire policies that take into account all reward components, allowing for a more balanced and versatile decision-making process.

Original languageEnglish
Pages (from-to)5174-5181
Number of pages8
JournalIEEE Robotics and Automation Letters
Volume9
Issue number6
DOIs
StatePublished - 1 Jun 2024

Keywords

  • Reinforcement learning
  • continual learning

ASJC Scopus subject areas

  • Mechanical Engineering
  • Control and Optimization
  • Artificial Intelligence
  • Human-Computer Interaction
  • Control and Systems Engineering
  • Computer Vision and Pattern Recognition
  • Biomedical Engineering
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'Adaptive Curriculum Learning With Successor Features for Imbalanced Compositional Reward Functions'. Together they form a unique fingerprint.

Cite this