LOCI: Spatial Linear Memory for Streaming World Models
PhysVista: Benchmarking Physical Intelligence in VLMs via a Perception-Reasoning-Assessment Loop
I Have a Stream: Making Self-Supervised Learning Work on Continuous Video
GAZEleak: Passcode Inference Against Eye-tracking XR Devices Through External Observation
TrackEverything: Long Horizon Dense Tracking via De-Duplicating 3D Scene Representations
SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models
