الملخص
World-action models (WAMs) have emerged as a promising paradigm for robot manipulation by jointly modeling future visual dynamics and robot actions. However, existing WAMs are trained predominantly on successful trajectories, making them prone to failure when real-world execution diverges from the learned dynamics. This issue is amplified in autoregressive WAMs, where execution errors become part of the causal history and continue to influence subsequent predictions. To this end, we introduce \method{}, a training-free framework that reformulates failure recovery as \emph{test-time scaling over causal histories}. This formulation decomposes recovery into three coupled decisions: \emph{when} to revise the causal history, \emph{where} to recover a reliable history prefix, and \emph{which} history configuration best supports subsequent execution. Specifically, \method{} realizes these decisions through three stages: 1) \textbf{Progress-Aware Recovery Trigger} detects persistent non-progress and triggers recovery only when the current execution state permits intervention; 2) \textbf{History-Prefix Recovery} identifies the unreliable history suffix, retrieves a historical anchor matching the current physical state, and reconstructs the causal KV state from the retained prefix while conditioning on the latest real observation; and 3) \textbf{Hypothesis Verification} compares the future continuations induced by complete-history, recovered-prefix, and full-reset hypotheses, and commits the best-supported hypothesis. Experiments in both simulated and real-world manipulation settings demonstrate consistent improvements in task success, while ablations confirm the contribution of each recovery stage.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Li, L., Chen, L., Kwunhang, Wong, Lei, J., Jin, S., Du, S., Zhang, C., Ma, S., Zhang, W., Xiao, J., Kwang-Ting, & Cheng (2026). Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models. https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models
MLA 9
Li, Lin, et al. "Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models." https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models.
شيكاغو (المؤلف–التاريخ)
Li, Lin, Long Chen, Kwunhang, Wong, Jiaming Lei, Song Jin, Shucheng Du, Chuhan Zhang, Songchen Ma, Weihao Zhang, Jun Xiao, Kwang-Ting, and Cheng. 2026. "Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models." https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models.
هارفارد
Li, L., Chen, L., Kwunhang, Wong, Lei, J., Jin, S., Du, S., Zhang, C., Ma, S., Zhang, W., Xiao, J., Kwang-Ting and Cheng (2026) 'Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models', Available at: https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models.
فانكوفر
Li L, Chen L, Kwunhang, Wong, Lei J, Jin S, et al. Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models. https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models
IEEE
L. Li, L. Chen, Kwunhang, Wong, J. Lei, S. Jin, S. Du, C. Zhang, S. Ma, W. Zhang, J. Xiao, Kwang-Ting, and Cheng, "Causal-History Test-Time Scaling for Failure Recovery in Autoregressive World-Action Models," https://omanscience.com/ar/articles/causal-history-test-time-scaling-for-failure-recovery-in-autoregressive-world-action-models.