← 研踪 PaperTrailarXiv:2611.25671 · v1
LATER: Learning Action-Conditioned Temporal Evidence Retrieval for Predictive World Models
首次提交 2026-11-19 15:48
版本更新 2026-11-19 15:48
雷达观测 2026-11-20 19:58
摘要
Early interactions can reveal physical properties that later observations no longer show. These properties can still gov- ern the effects of future actions. Inspired by the distinction between retention and selective readout in working mem- ory and cognitive science, we introduce LATER, an action- conditioned world model built on CF-JEPA. LATER stores completed real transitions in partitioned memories that remain fixed during candidate rollouts. A response bridge queries these memories using the current state and candidate action, while multi-cut auxiliary supervision trains the memory writ- ers. On distracted Two-Stage PushT, LATER raises terminal success from 15.5% to 25.0% over reconstructed CF-JEPA. In a separate comparison across six models and two scratch- training seeds, LATER achieves the lowest mean physical readout error in both clean and distracted views, reducing along-push mean absolute error (MAE) under distractors by 20.2% relative to CF-JEPA. Controlled history interventions show that later predictions depend on earlier physical evi- dence even when query observations and actions are fixed, with larger effects during pushing than waiting.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-19 15:48 · 最新观测版本
LATER: Learning Action-Conditioned Temporal Evidence Retrieval for Predictive World Models
Early interactions can reveal physical properties that later observations no longer show. These properties can still gov- ern the effects of future actions. Inspired by the distinction between retention and selective readout in working mem- ory and cognitive science, we introduce LATER, an action- conditioned world model built on CF-JEPA. LATER stores completed real transitions in partitioned memories that remain fixed during candidate rollouts. A response bridge queries these memories using the current state and candidate action, while multi-cut auxiliary supervision trains the memory writ- ers. On distracted Two-Stage PushT, LATER raises terminal success from 15.5% to 25.0% over reconstructed CF-JEPA. In a separate comparison across six models and two scratch- training seeds, LATER achieves the lowest mean physical readout error in both clean and distracted views, reducing along-push mean absolute error (MAE) under distractors by 20.2% relative to CF-JEPA. Controlled history interventions show that later predictions depend on earlier physical evi- dence even when query observations and actions are fixed, with larger effects during pushing than waiting.
查看此版本解读 →查看此版本原文 ↗