FaST method proposed for factored synthesis of transitions in offline reinforcement learning
A new arXiv paper introduces FaST, a factored synthesis approach for generating transitions in offline reinforcement learning, addressing limits of diffusion-based synthetic data generators that model fixed logged datasets.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-11-11 05:00 · arXiv cs.LG
FaST: Factored Synthesis of Transitions for Offline Reinforcement Learning