Simulation-based inference with neural posterior estimation (NPE) often yields overconfident and unreliable posteriors under limited simulation budgets. To address this, we propose DRO-NPE, a distributionally robust approach that replaces the standard NPE objective with a worst-case loss over a Wasserstein ambiguity set. We introduce KL-based metrics for miscoverage and miscalibration, and use these to show that the DRO-NPE objective controls overfitting and reduces posterior overconfidence. Our method is tractable, parallelisable, and readily integrates with standard normalising flows. Across benchmark SBI tasks, DRO-NPE consistently improves coverage and calibration, while narrowing the gap between empirical and population NPE loss, leading to more reliable inference in low-simulation regimes.

翻译：基于神经后验估计（NPE）的仿真推理在有限仿真预算下往往产生过度自信且不可靠的后验分布。为解决这一问题，我们提出 DRO-NPE，一种通过 Wasserstein 模糊集上的最坏情况损失替代标准 NPE 目标的分布鲁棒方法。我们引入基于 KL 散度的误覆盖与误校准指标，并利用这些指标表明 DRO-NPE 目标能有效控制过拟合并减少后验过度自信。该方法具有可计算性、可并行化，并能与标准归一化流直接集成。在基准 SBI 任务中，DRO-NPE 持续提升覆盖率和校准质量，同时缩小经验 NPE 损失与总体 NPE 损失之间的差距，从而在低仿真场景下实现更可靠的推理。