发表机构
School of Computing and Augmented Intelligence Arizona State University（计算与增强智能学院亚利桑那州立大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对概率电路学习易过拟合和泛化能力弱的问题，提出PeTeR这一无需数据的训练后框架，可使预训练概率电路抵御分布变化，实证评估显示该框架能有效增强模型对随机和对抗性扰动的鲁棒性，性能良好。
AI 中文摘要
概率电路（PCs）能对复杂联合分布建模，并支持对许多推理查询进行精确高效计算。然而，标准的基于似然的PC学习在面对数据噪声、小样本量或分布变化时易过拟合且泛化能力弱。可通过分布鲁棒优化缓解，但当前方法限于在此框架下从头训练模型。我们提出PeTeR，一个无需数据的训练后框架，能使预训练PCs抵御分布变化而无需从头再训练。在多个密度估计基准上的实证评估表明，PeTeR能有效增强基线模型抵御随机和对抗性扰动的能力，性能优于或可媲美依赖数据的鲁棒学习基线。
英文摘要
Probabilistic circuits (PCs) can model complex joint distributions while supporting exact and efficient computation of many inference queries. However, standard likelihood-based PC learning is vulnerable to overfitting and fragile generalization when confronted with data noise, small sample sizes, or distribution shifts. This can be mitigated using distributionally-robust optimization which consider worst-case distributions within a Wasserstein ball of the empirical distribution, but current methods are limited to training a model from scratch in this framework. Instead, we propose PeTeR: a novel, data-free post-training framework designed to robustify pre-trained PCs against distribution shifts without retraining from scratch. Empirical evaluations across multiple density estimation benchmarks demonstrate that PeTeR effectively robustifies baseline models against both random and adversarial perturbations, achieving competitive or superior performance to data-dependent robust learning baselines.
