Liuyi Wang

Real-time embodied navigation no longer requires slow autoregressive MLLM rollouts at every timestep: compressing candidate waypoints into discriminative semantic tokens enables sub-second decisions on continuous navigation benchmarks with zero loss in task success.

Achieving $0.1\text{m}$ spatial resolution with real-time inference, GaussianSeed redefines the efficiency-quality balance in 3D occupancy prediction.

Memory-augmented pose estimation can dramatically improve generalization across diverse object instances, outperforming traditional methods by leveraging accumulated geometric knowledge.

TACO achieves state-of-the-art performance in open-vocabulary video recognition by preserving out-of-distribution alignment, challenging the conventional trade-off between generalization and specialization.