Da-Cheng Tao

A novel tri-stream transformer architecture unlocks synergistic control by enabling a dedicated action expert to selectively fuse semantic and dynamics information from independent VLM and World Model streams.

This framework surveys manipulation, navigation, locomotion, autonomous driving, and general embodied learning, tracing technical progressions, clarifying capability requirements, and examining datasets, benchmarks, and evaluation protocols.

Online submodular optimization cannot match the classic $1-1/e$ guarantee under stable updates: beating $2-\sqrt{2} \approx 0.5858$ provably triggers either an exponential query explosion or catastrophic $\Omega(k)$ recourse per step.

A simple physical object can hijack an agent's internal imagination rollouts, trapping its downstream policy in an adversarial trajectory long after the trigger leaves the scene.

Heavy-tailed designs can lead to a dramatic increase in sample complexity, challenging established norms in statistical recovery.

Nuisance interpolation in feature priming can lead to significant underperformance in online regression, revealing a fundamental limitation in existing regret guarantees.

MSCNet not only reconstructs missing MRI sequences but also enhances lesion fidelity, outperforming traditional methods in clinical settings.

SafeCA slashes jailbreak success rates by 20% while adding virtually no latency, revolutionizing defenses for text-to-video models.

Combodied Agents redefine AI by shifting focus from task completion to fostering sustained human well-being through adaptive, user-centered support.

ACE-Data-0 reveals that existing models struggle significantly with complex interactions, exposing critical gaps in embodied AI performance.