发表机构
Texas A&M University（德克萨斯农工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
ActionSplice通过反事实状态传输框架，在冻结世界模型下实现飞行中动作编辑，显著降低回滚延迟和感知误差，提升交互式采样效率。
AI 中文摘要
分块自回归视频世界模型通常将每个生成的块与一个动作条件化。因此，在采样过程中接收到的动作必须等待下一个块，将未来的求解器评估条件化于在前一个动作下产生的状态，或触发回滚以重复已完成的评估。我们引入了ActionSplice，一个推理框架，将这一问题表述为反事实状态传输（CST）。一个轻量级校正器将中断的骨干原生表示传输到由修订动作在相同求解器步骤下诱导的匹配状态。世界模型和采样器保持冻结，采样无需重放已完成的评估即可恢复。重定向变体CST_R更新整个活动块，而时间拼接变体CST_T保留时间前缀并仅更新后缀。在minWM-Wan Action2V和HY-WM1.5上，相对于直接条件交换，CST_R将回滚相对LPIPS分别降低了61.5%和75.9%。CST_T分别将后缀LPIPS降低了56.1%和77.5%，同时相对于等待提供了2.73倍和1.69倍的像素就绪加速。在HY-WorldPlay协议下，CST_R相对于原始rollout获得了25.66 dB的PSNR、0.6902的SSIM和0.1337的LPIPS。
英文摘要
Chunk-autoregressive video world models typically generate each chunk under one action. When an action changes during sampling, waiting until the next chunk delays the response, while directly switching the conditioning leaves the intermediate solver state shaped by the previous action. Restarting sampling under the revised action avoids this mismatch but repeats completed computation. We introduce \emph{ActionSplice}, which edits the interrupted state through Counterfactual State Transport (CST). A lightweight corrector moves the interrupted solver state toward the matched counterfactual solver state induced by the revised action at the same solver step. The world model and sampler remain frozen, and sampling resumes without replaying completed evaluations. $\mathrm{CST}_{R}$ retargets the entire active chunk, while $\mathrm{CST}_{T}$ preserves the temporal prefix at the intervention step and corrects only the suffix. On minWM--Wan Action2V and HY-WM1.5, we measure fidelity against matched rollback references. Relative to condition swapping, $\mathrm{CST}_{R}$ reduces LPIPS by 61.5\% and 75.9\%, and $\mathrm{CST}_{T}$ reduces suffix LPIPS by 56.1\% and 77.5\%, respectively.
Comments16 pages, 6 figures. Project page: https://pardistaghavi.github.io/actionsplice-website/
