Newest papers

Knowing When to Stop Thinking: A Dissection of Expected Free Energy as a Test-Time Compute Controller for Latent World-Model Planning

Tong Zhang, Zhiyuan Shi, Yun Peng, Tao Xie · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28011 · cited by 0 works (OpenAlex) · open-access PDF

Chinese

使用学术世界模型的规划者可以改善其预测结果,同时选择更差的行动。我们研究预期免费能量(EFE)可以帮助决定在计划上花费多少计算。以EFE启发的目标距离与斯托哈斯特预测之间的分歧相结合。我们单独测试候选人分数,停止和在线预测更新。当精英成本停留时停止减少了成功平台深度的任务的精细化,平均成功接近固定深度参照。在较大的测试重量上,奖励分歧降低了冻结模型的成功。惩罚它有助于某些任务设置和重量设置,但伤害其他人。当训练数据忽略到一个需要达到目标的区域时,在线更新恢复了结果的一部分。在优化模型错误时,丢失的成本规则可以进一步失

Machine translation (Meta M2M100); the English original follows.

Abstract

A planner using a learned world model can improve its predicted outcome while choosing a worse action. We study when expected free energy (EFE) can help decide how much computation to spend on a plan. An EFE-inspired objective combines goal distance with disagreement among stochastic predictions. We test candidate scoring, stopping, and online predictor updates separately. Stopping when elite cost stalls reduces refinement on tasks whose success plateaus with depth, with mean success close to fixed-depth references. At the larger tested weights, rewarding disagreement lowers success with a frozen model. Penalizing it helps at some task and weight settings but harms others. When training data omit a region needed to reach the goal, online updates recover part of the resulting loss. The tested cost-stall rule can fail when further optimization exploits model errors, because predicted cost keeps falling after task success peaks. Cost progress can guide computation on a plateau, but protecting performance after a peak requires information about the consequences of the chosen actions.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Task (project management), Computer science, Computation, Exploit, Outcome (game theory), Energy (signal processing)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange