RAVEN: 实时自回归视频外推与一致性模型GRPO
RAVEN: Real-time Autoregressive Video Extrapolation with Consistency-model GRPO
AI总结：
RAVEN通过将自回归视频扩散模型的训练与测试过程结合，提升长时视频生成质量。CM-GRPO方法优化一致性采样，提高生成效果。
摘要：
因果自回归视频扩散模型通过外推未来片段实现实时流式生成。从高保真双向教师模型中蒸馏生成器可获得竞争性少步模型，但训练时的历史分布与推理时的分布差异限制了长时生成质量。我们提出实时自回归视频外推网络（RAVEN），一种训练时测试框架，将每个自回归滚动转换为清洁历史端点和噪声去噪状态的交错序列。此方法使训练注意力与推理时的外推对齐，并允许下游片段损失监督未来预测依赖的历史表示。我们进一步提出一致性模型组相对策略优化（CM-GRPO），将一致性采样步骤重新表述为条件高斯转移，并直接对这一核应用在线强化学习（RL），避免了先前流模型RL方法中采用的欧拉-马尔蒂娜辅助过程。实验表明，RAVEN在质量、语义和动态度评估中超越了最近的因果视频蒸馏基线，且CM-GRPO与RAVEN结合时进一步提升了性能。
英文摘要：
Causal autoregressive video diffusion models support real-time streaming generation by extrapolating future chunks from previously generated content. Distilling such generators from high-fidelity bidirectional teachers yields competitive few-step models, yet a persistent gap between the history distributions encountered during training and those arising at inference constrains generation quality over long horizons. We introduce the Real-time Autoregressive Video Extrapolation Network (RAVEN), a training-time test framework that repacks each self rollout into an interleaved sequence of clean historical endpoints and noisy denoising states. This formulation aligns training attention with inference-time extrapolation and allows downstream chunk losses to supervise the history representations on which future predictions depend. We further propose Consistency-model Group Relative Policy Optimization (CM-GRPO), which reformulates a consistency sampling step as a conditional Gaussian transition and applies online Reinforcement Learning (RL) directly to this kernel, avoiding the Euler-Maruyama auxiliary process adopted in prior flow-model RL formulations. Experiments demonstrate that RAVEN surpasses recent causal video distillation baselines across quality, semantic, and dynamic degree evaluations, and that CM-GRPO provides further gains when combined with RAVEN.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
