arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2607.03039cs.LGcond-mat.stat-mech
收藏

动态伊辛模型中的分布外神经推理

Out-of-distribution Neural Inference in Dynamical Ising Models

Center for Quantum Physics and Intelligent Sciences, Department of Physics, Capital Normal University(首都师范大学物理学院量子物理与智能科学中心)

Yuan-Bin Zhu, Shuang Qiao, Shi-Ju Ran

AI总结:

研究通过从格劳伯磁化轨迹重建动力学伊辛模型的相互作用图这一反问题,探讨神经网络分布外性能是否反映可转移物理规则学习。发现不同架构训练有不同统计策略,指出分布内高精度和分布外鲁棒性不一定意味着学到动力学与结构规则。

摘要:

神经网络越来越多地用于从动态观测中推断隐藏的物理结构,但其分布外性能是否反映可转移的物理规则学习仍不清楚。我们在一个受控逆问题中解决这个问题:从格劳伯磁化轨迹重建动力学伊辛模型的相互作用图。我们发现,在拓扑和温度变化下,数据驱动的训练会产生不同且可重复的统计策略。边缘种群诊断表明,基于Transformer的模型倾向于保持训练集的链接密度,而卷积模型可以通过利用多数无链接类,向稀疏或无链接预测收敛,这些预测在分布外似乎是稳定的。因此,高分布内准确率和明显的分布外鲁棒性不一定意味着学习到了动力学与结构规则。相反,神经重建可以由依赖于架构的统计先验来控制。我们的结果确定了物理逆问题中标准数据驱动学习的具体失败模式,并为机器学习辅助的科学发现提出了规则指导原则。

英文摘要:

Neural networks are increasingly used to infer hidden physical structure from dynamical observations, yet it remains unclear whether their out-of-distribution performance reflects transferable physical rule learning. We address this question in a controlled inverse problem: reconstructing interaction graphs of a kinetic Ising model from Glauber magnetization trajectories. Across convolutional, graph, Transformer, and hybrid architectures, we find that data-driven training produces distinct and reproducible statistical strategies under topology and temperature shifts. Edge-population diagnostics reveal that Transformer-based models tend to preserve the link density of the training ensemble, whereas convolutional models can collapse toward sparse- or no-link predictions that appear out-of-distribution stable by exploiting the majority no-link class. Thus, high in-distribution accuracy and apparent out-of-distribution robustness do not necessarily imply a learned dynamics-to-structure rule. Instead, neural reconstruction can be governed by architecture-dependent statistical priors. Our results identify a concrete failure mode of standard data-driven learning in physical inverse problems and motivate rule-guided principles for machine-learning-assisted scientific discovery.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑