BAgger：用于缓解自回归视频扩散模型漂移的反向聚合方法
BAgger: Backwards Aggregation for Mitigating Drift in Autoregressive Video Diffusion Models
针对自回归视频扩散模型的曝光偏差问题，BAgger 用模型自身 rollout 构造纠错轨迹，并以标准 score/flow matching 目标训练，无需少步蒸馏或长链 BPTT；做长视频自回归生成或世界模型时可作为缓解误差累积的通用训练策略。
原文依据 constructs corrective trajectories from the model's own rollouts, teaching it to recover from its mistakes
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论自回归视频模型通过下一帧预测进行世界建模颇具前景，但存在曝光偏差问题：在干净上下文上训练与在自生成帧上推理之间存在不匹配，导致误差累积、质量随时间漂移。我们提出反向聚合（BAgger），一种自监督方案，从模型自身的 rollout 中构造纠错轨迹，教会模型从错误中恢复。与依赖少步蒸馏和分布匹配损失的先前方法（可能损害质量与多样性）不同，BAgger 使用标准的 score 或流匹配目标进行训练，避免大型教师模型和长时间链式时间反向传播。我们在因果扩散 Transformer 上实例化 BAgger，并在文本到视频、视频延展和多提示生成上评估，观察到更稳定的长时程运动以及更好的视觉一致性，漂移减少。
英文原文摘要 941 字符
Autoregressive video models are promising for world modeling via next-frame prediction, but they suffer from exposure bias: a mismatch between training on clean contexts and inference on self-generated frames, causing errors to compound and quality to drift over time. We introduce Backwards Aggregation (BAgger), a self-supervised scheme that constructs corrective trajectories from the model's own rollouts, teaching it to recover from its mistakes. Unlike prior approaches that rely on few-step distillation and distribution-matching losses --which can hurt quality and diversity -- BAgger trains with standard score or flow matching objectives, avoiding large teachers and long-chain backpropagation through time. We instantiate BAgger on causal diffusion transformers and evaluate on text-to-video, video extension, and multi-prompt generation, observing more stable long-horizon motion and better visual consistency with reduced drift.