Minjun Kang

Asynchronous OPD can boost training throughput significantly while managing the challenges of stale data, transforming the efficiency of large language model fine-tuning.

Event cameras unlock 6D pose tracking of novel objects at 120+ FPS, even with rapid motion, by fusing sparse event streams with depth in a way that generalizes zero-shot from synthetic training.

By adapting diffusion features in 3D Gaussian space, GeoNVS achieves state-of-the-art novel view synthesis with significantly improved geometric fidelity and camera control compared to existing video diffusion models.