Lingdong Kong

Driving models can achieve near-perfect performance with just one denoising step by learning to predict visual motion directly, outperforming multi-step prediction models and achieving 3.29x inference speed.

Autonomous driving VLAs can be made conflict-aware by dreaming world models that strictly follow navigation signals, revealing unsafe instruction scenarios.

VLMs can be taught to *use* latent geometry by training them to read continuous geometric tokens, rather than just reconstructing geometry from RGB.

Action-class miscalibration is a hidden failure mode that can significantly skew performance evaluations of LLMs in multi-turn tool-calling tasks.

On-policy self-distillation can boost diffusion model performance by up to 44% while slashing training time by over 60%.

Ten teams achieved groundbreaking advancements in low-light image enhancement, with improvements that redefine the state-of-the-art in burst photography.