Jingyuan Wang

Urban spatial optimization can be solved far more effectively by framing it as an interactive code-and-eval sandbox with decomposed build-and-refine RL than by relying on rigid, task-specific RL solvers.

Refined annotations can boost PE segmentation performance more than changes in model training, challenging assumptions about model superiority.

DEPT achieves superior retrieval quality by preserving document embeddings while adapting query expansions, revealing a new synergy between these traditionally separate tasks.