Video Models Can Reason with Verifiable Rewards by Tinghui Zhu
Abstract: Video diffusion models are emerging as general-purpose visual foundation models. Yet, the power of such generalization can be hindered by a reasoning ability that lags behind large language models. The main gap is the mismatch between perceptual plausibility and objective correctness. SFT on ground-truth solution videos can teach the model the visual form of valid trajectories, yet it does not directly optimize the correctness of sampled outputs. As a result, models may imitate solution-like patterns while failing to satisfy the underlying rules that make those solutions valid. In this talk, I will share our recent work VideoRLVR, a framework that trains video diffusion models with reinforcement learning from verifiable rewards. I will also discuss lessons from our failed attempts and how they lead to a practical recipe for moving video models from visual imitation toward verifiable reasoning.
Speaker: Tinghui Zhu — Ph.D. student at UC Davis, working on reasoning and agents across different modalities.
Website: https://journal.video-reason.com/ Register on this page to receive the Zoom join link (sent automatically with your confirmation).
Subscribe to our mailing list for weekly talk announcements: Open Google Form