X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding Paper • 2606.02482 • Published Jun 1 • 32
CogOmniControl: Reasoning-Driven Controllable Video Generation via Creative Intent Cognition Paper • 2605.19995 • Published May 19 • 34
From Web to Pixels: Bringing Agentic Search into Visual Perception Paper • 2605.12497 • Published May 12 • 13