📄 中文摘要
以往关于反事实识别的完整性结果主要局限于观察性或干预性分布的情境（Pearl因果层次的第一层和第二层），因为通常认为无法从反事实分布中获取数据，而反事实分布属于第三层。然而，最近的研究（Raghavan & Bareinboim, 2025）正式描述了一类可以通过实验方法直接估计的反事实分布，称为反事实可实现性。这一进展引发了一个重要问题：在获得这些反事实分布的情况下，哪些额外的反事实量现在变得可识别？
来自反事实数据的因果识别：完整性和界限结果
📄 English Summary
Causal Identification from Counterfactual Data: Completeness and Bounding Results
Previous work on completeness results for counterfactual identification has been limited to settings where input data comes from observational or interventional distributions (Layers 1 and 2 of Pearl's Causal Hierarchy), as it was generally deemed impossible to obtain data from counterfactual distributions, which belong to Layer 3. However, recent research by Raghavan & Bareinboim (2025) has formally characterized a family of counterfactual distributions that can be directly estimated through experimental methods, a concept they refer to as counterfactual realizability. This advancement raises an important question regarding what additional counterfactual quantities now become identifiable with this new access to certain counterfactual distributions.