Keep It Simple: Multi-Key Episodic Memory Retrieval for Ultra-Long Video Understanding Paper • 2608.07663 • Published 25 days ago • 23
Scenes as Objects, Not Primitives: Instance-Structured 3D Tokenization from Unposed Views Paper • 2606.29513 • Published Jun 28 • 53
Where Did It Go Wrong? Process-Level Evaluation of Web Agents with Semantic State Tracking Paper • 2606.15673 • Published Apr 8 • 13