Nanye Ma

Few-step diffusion models can achieve state-of-the-art 4-NFE generation with simple on-policy velocity regression and a novel residual proxy, matching multi-step teacher performance without complex auxiliary models.

MLLMs are failing to visually track events in videos, performing only modestly above baseline despite strong results on other benchmarks.

Generate minute-long videos with compelling narrative structure and local realism, even with limited long-form training data, by cleverly combining supervised flow matching for global coherence with mode-seeking alignment to a short-video teacher for local fidelity.