提出新基准R2M-Bench，用相对一致性评估视频世界模型的重访记忆能力。
R2M-Bench: Evaluating Revisit Memory via Relative Consistency in Interactive Video World Models
- 通过同次回放中的对比对，计算重访优势与归一化记忆比。
- 在300个场景中验证，该方法与人类判断相关性达0.547。
- 可有效区分记忆行为与画面稳定性的混淆信号，适合模型开发者使用。
首次访问与返回帧高度相似并不必然说明视频世界模型记住了场景，中间生成过程可能几乎无变化。这种模糊性使绝对重访分数受渲染稳定性、重复内容和运动失败的影响。本文提出 extbf{R2M-Bench}（相对重访记忆基准），基于可观测的重访选择性一致性进行评估。对于每个检测到的返回事件，R2M-Bench将重访对与同次回放中的两个对照组比较：一个时间间隔匹配的非重访对（衡量通用时间稳定性），以及一个短距离对（估计短期一致性）。由此得出 extit{MemoryGain}（MG）——重访相对于时间基线的优势，以及 extit{Normalized Memory Ratio}（NMR）——将该优势归一化至短时-基线动态范围。该基准结合100个参考场景与三条离开-返回轨迹，共形成300个实例，评估外观保真度、场景与物体身份、局部几何及持久状态。在七种动作条件视频世界模型中，总体NMR与人类一致性判断的相关系数为Spearman's $ρ=0.547$（95% CI $[0.45,0.63]$）。其模型内与生成运动的相关性为0.072，远低于原始重访相似性的0.207，表明相对校准显著抑制了慢动作捷径。DreamX-World-Memo在所评模型中取得最高总体NMR。结果共同支持：同次回放的相对校准是区分重访特异性一致性与通用时间稳定性的实用方法。
原文摘要 · Abstract (English)
High similarity between first-visit and return frames does not necessarily show that a video world model remembered the scene; the intervening rollout may simply have changed very little. This ambiguity makes absolute revisit scores sensitive to rendering stability, repetitive content, and failed motion. We introduce \emph{R2M-Bench} (\textbf{R}elative \textbf{R}evisit \textbf{M}emory Benchmark), a benchmark of observable revisit-selective consistency. For every detected return, R2M-Bench compares the revisit pair with two controls from the same rollout: a gap-matched non-revisit pair that measures generic temporal stability and a short-range pair that estimates short-horizon consistency. These comparisons produce \emph{MemoryGain} (MG), the revisit advantage over the temporal baseline, and the \emph{Normalized Memory Ratio} (NMR), which normalizes this advantage by the short-to-baseline dynamic range. R2M-Bench combines 100 reference scenes with three leave-and-return trajectories to form 300 instances and evaluates appearance fidelity, scene and object identity, local geometry, and persistent state. Across seven action-conditioned video world models, Overall NMR correlates with human consistency judgments at Spearman's $ρ=0.547$ (95\% CI $[0.45,0.63]$). Its within-model correlation magnitude with generated motion is $0.072$, compared with $0.207$ for raw revisit similarity, indicating that relative calibration substantially reduces the slow-motion shortcut. DreamX-World-Memo achieves the highest Overall NMR among the evaluated video models. Together, these results support same-rollout relative calibration as a practical way to distinguish revisit-specific consistency from generic temporal stability.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。