让分散机器人共享未来状态预测，无需集中协调或标签。
One Future, Every Robot: Label-Efficient Collective-State Prediction with Decentralized JEPA
- 每台机器人基于自身观测和邻居信息，预测同一固定维度的未来隐状态。
- 相比原始未来重建，集体状态误差降低28.4%（分布内）至75.6%（拓扑变化）。
- 无需共享决策，适合多智能体协同任务，尤其标签稀缺场景。
分散式机器人常需对团队整体状态达成一致认知，但每台机器人仅能获取局部观测，且无法依赖中心估计或输出共识。我们探讨在该约束下，能否实现一致的集体状态预测。集体状态JEPA（CS-JEPA）训练每台机器人从自身历史与有限邻居消息中，预测相同的固定宽度未来隐状态，不使用一致性损失；部署时预测与计划永不聚合。在独立复现中，所有种子与评估分割下的预测一致性均提升。准确率同步改善，排除了所有机器人退化为单一预测的无意义解：相对于容量匹配的原始未来重建，集体状态误差在分布内降低28.4%，在拓扑与群组规模变化下降低64.4%至75.6%。无翻译与交叉预训练控制实验保持此联合效果；动作条件与刚体评估表明接收端局部表示支持独立决策。因此，共享隐未来状态可在无共识训练下对齐分散预测，同时保留有用且标签高效的表征。
原文摘要 · Abstract (English)
Decentralized robots often need a common view of what their team is becoming, even though each robot sees different evidence and cannot rely on a central estimate or output-level consensus. We ask whether compatible collective-state predictions can emerge under this constraint. Collective-State JEPA (CS-JEPA) trains every robot to predict the same fixed-width latent future from its own history and bounded neighbor messages, with no agreement loss; predictions and plans are never pooled at deployment. In a fresh independent replication, agreement improves for every seed and every evaluated split. Accuracy improves at the same time, ruling out the uninformative solution in which all robots merely collapse to one prediction: relative to capacity-matched raw-future reconstruction, collective-state error falls by 28.4 percent in distribution and by 64.4 to 75.6 percent under topology and swarm-size shift. Translation-free and crossed-pretraining controls preserve this joint result, while action-conditioned and rigid-body evaluations show that the receiver-local representation supports independent decisions. A shared latent future can therefore align decentralized predictions without consensus training while preserving useful, label-efficient information.
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。