Omri Uzan

Document optimization via RL boosts retrieval performance for smaller models

Researchers have developed a novel document optimization technique using reinforcement learning, specifically GRPO, to enhance retrieval quality. This method fine-tunes language models to transform documents into repres…