研究人员正在开发新方法来改进机器学习模型中的因果推断，特别是在处理未测量混淆因素时。一种方法，隐藏路径贡献（HPC），旨在区分通过合法模型路径产生正确答案的干预措施与利用隐藏路径的干预措施。另一个研究领域侧重于因果基础模型（CFMs），它们利用多样化的实验方案和观测数据来更准确地预测条件干预分布。研究还探讨了可用观测数据如何影响因果效应的识别，并引入了因果ID视图（CausalIDView）等基准来评估CFMs在不同数据视图下的性能。此外，还开发了近端平衡（Proximal Balancing）和SPICE-Net等方法，利用代理变量来处理未测量的混淆因素，为更鲁棒的因果效应估计提供理论保证和实用算法。
AI
arXiv:2609.39243v1 Announce Type: new Abstract: Causal interventions such as activation patching and distributed alignment search (DAS) are the main tool for making mechanistic claims about neural networks. Recent work showed that these interventions routinely push representation…
arXiv cs.LG
TIER_1English(EN)·Yuche Gao, Arik Reuter, Siyuan Guo, Anish Dhir, Bernhard Sch\"olkopf, Adrian Weller·
arXiv:2609.39523v1 Announce Type: new Abstract: Causal foundation models (CFMs) amortise causal inference over priors of synthetic structural causal models (SCMs), predicting the effect of an experiment on a specific variable. However, observational data alone may leave multiple …
arXiv cs.LG
TIER_1English(EN)·Heejin Jung, Gyeongdeok Seo, Hoyoon Byun, Joseph Lee, Kyungwoo Song·
arXiv:2609.36881v1 Announce Type: new Abstract: Causal foundation models (CFMs) pre-trained on data generated from various structural causal models (SCMs) have been proposed for estimating causal effects from observational data. However, differences in pre-training environments a…
arXiv:2609.40051v1 Announce Type: new Abstract: Estimating causal effects from observational data is central to science and policy, but the effects are not identified when confounders are unmeasured. Proximal causal inference addresses this problem with proxies of the unmeasured …
arXiv:2008.13651v4 Announce Type: replace-cross Abstract: This paper develops a causal inference method for treatment effects models with noisily measured confounders. The key feature is that a large number of noisy proxies are available and linked with the underlying latent conf…
arXiv stat.ML
TIER_1English(EN)·Silvan Vollmer, Niklas Pfister, Sebastian Weichwald·
arXiv:2604.09135v2 Announce Type: replace Abstract: Unobserved confounding is a key challenge when estimating causal effects from a treatment on an outcome. In this work, we assume that we observe a single, potentially multi-dimensional proxy variable of the unobserved confounder…
