OCIS · Fall 2023¶
- 共 13 场 · 11 篇精读
本季导览¶
自动生成：归纳本季主线与值得先看的几场，不打分、不排名。
OCIS Fall 2023 的 13 场报告可归纳为四条主线：因果参数的实质解释与识别陷阱（Stensrud & Sarvet、Miles 等）、面板数据与干扰下的因果推断（Agarwal、Hudgens & Lee）、混杂与选择偏差的图形化调整（Mathur、Guo）、以及潜变量模型与干预泛化的可识别性（Gu、Silva）。此外，贝叶斯自适应估计（Gelman）和多因素观测研究的加权方法（Yu & Ding）构成两条独立但互补的方法论线索。
因果参数的实质解释与识别陷阱是贯穿全季的反思性主线。Stensrud & Sarvet 系统诊断了“身份滑移”——统计上可识别的参数被错误解释为研究者心中理想的因果问题，并呼吁回归问题驱动传统。Miles 的报告（含 Robins 与 Richardson 的讨论）则聚焦中介分析，指出“随机干预自然间接效应”虽放松了识别假设，却可能不捕捉实质中介，并对比了可分离效应框架。两条报告共同追问：当统计便利性与实质解释冲突时，如何避免参数滑移？Gelman 的报告虽不直接讨论解释，但其贝叶斯分层模型在“无偏但高方差”与“有偏但低方差”估计量间的自适应权衡，实质上也在回应同一张力——如何根据数据决定偏倚-方差折中，而非默认使用保守估计量。
面板数据与干扰下的因果推断是方法论推进最密集的主线。Agarwal 将合成控制法扩展到存在时间与空间溢出的面板数据，利用低秩潜因子模型同时处理两类溢出，并给出识别条件与估计量。Hudgens & Lee 则聚焦集群干扰，针对“Type B”政策（集群内个体处理概率固定）提出半参数有效估计量，并推广到随机政策效应。两条报告共享“干扰”主题，但切入角度不同：Agarwal 强调潜因子结构下的反事实预测，Hudgens & Lee 则侧重集群内干扰的随机化与效率界。Yu & Ding 的析因研究加权方法虽不直接处理干扰，但其对多因子主效应与交互效应的统一加权框架，在面板数据中也可用于处理多时间点干预。
混杂与选择偏差的图形化调整提供了两条互补的实用工具。Mathur 提出“选择偏差共同原因原则”，利用 DAG/SWIG 给出协变量调整消除选择偏差的充分条件，无需完整指定因果结构。Guo 则针对因果图不完全已知的场景，提出迭代图扩展程序，通过局部结构知识逐步构建充分调整集。两条报告都试图降低图形方法的应用门槛：Mathur 聚焦选择偏差（如 collider bias），Guo 聚焦混杂控制，但均强调“局部知识”而非全图假设。
潜变量模型与干预泛化的可识别性是理论深度最高的主线。Gu 建立了离散潜变量深度生成模型的严格可识别性，利用稀疏图结构条件保证潜变量个数、取值与依赖关系的唯一恢复。Silva 则从因子图模型出发，处理干预泛化问题——如何从已知干预 regime 预测新干预组合下的分布，并给出基于因子图的可识别条件。两条报告都涉及“结构可识别性”这一核心概念，但 Gu 聚焦潜变量恢复，Silva 聚焦干预分布预测。
若想快速把握全季脉络，建议按以下入口观看：因果解释与识别陷阱：先看 Stensrud & Sarvet 打底，再看 Miles 的中介分析讨论（含 Robins 与 Richardson 的讨论）进阶。面板数据与干扰：先看 Hudgens & Lee 理解集群干扰下的半参数效率，再看 Agarwal 处理时空溢出的合成控制扩展。图形化调整：先看 Guo 的迭代图扩展（混杂控制），再看 Mathur 的选择偏差共同原因原则（选择偏差）。可识别性：先看 Gu 的离散潜变量模型，再看 Silva 的因子图干预泛化。Gelman 的贝叶斯自适应估计和 Yu & Ding 的析因加权方法可分别作为“偏倚-方差权衡”和“多因子观测研究”的独立入口。
报告列表¶
Interpretational errors in causal inference and how to avoid them¶
讲者: Mats Stensrud, Aaron Sarvet · 讨论人: Kerollos Wanis and Vanessa Didelez . Q&A moderators: Lan Wen · 2023-12-12
链接：视频 · 幻灯片
摘要
Pioneering works in causal inference were explicitly grounded in practical disciplines, aiming at formalizing real questions with mathematical definitions. Now, causal inference methods provide an architecture that profoundly regulates what practical questions get asked and how they get answered. Here we consider subtly different approaches for causal inference research, and their implications for theory development and practice. In this process, we formalize an interpretational error that is in…Flexible modeling of adaptive treatment strategies for censored outcomes¶
讲者: Erica Moodie · 讨论人: Yu Cheng , Peter Thall · 2023-12-05
链接：视频 · 幻灯片
摘要
To achieve the goal of providing the optimal care to each patients, physicians must customize treatments. Making decisions at multiple stages as a disease progresses can be formalized as an adaptive treatment strategy (ATS). To be able to recommend an optimal treatment, an understanding of the causal effect of treatment is required. In this talk, I will discuss an extension of a general Bayesian machine learning framework for the popular estimation approach of Q-learning, adapted to censored out…Identifiable Deep Generative Models for Rich Data Types with Discrete Latent Layers¶
讲者: Yuqi Gu · 讨论人: Qingyuan Zhao · 2023-11-28
链接：视频 · 幻灯片
Identifiable Deep Generative Models for Rich Data Types with Discrete Latent Layers （暂无精读）¶
讲者: Yuqi Gu, - D, iscussant: Qingyuan Zhao · 2023-11-28
链接：视频 · 幻灯片
摘要
We propose a class of identifiable deep generative models for very flexible data types. The key features of the proposed models include (a) discrete latent layers and (b) a shrinking pyramid- or ladder-shaped deep architecture. We establish model identifiability by developing transparent conditions on the sparsity structure of the deep generative graph. The proposed identifiability conditions can ensure estimation consistency in both the Bayesian and frequentist senses. As an illustration, we co…A common-cause principle for eliminating selection bias in causal estimands through covariate adjustment¶
讲者: Maya Mathur · 讨论人: Eric Tchetgen Tchetgen and Nan Laird [new format] · 2023-11-14
链接：视频
摘要
Average treatment effects (ATEs) may be subject to selection bias when they are estimated among only a non-representative subset of the target population. Selection bias can sometimes be eliminated by conditioning on a “sufficient adjustment set” of covariates, even for some forms of missingness not at random (MNAR). Without requiring full specification of the causal structure, we consider sufficient adjustment sets to allow nonparametric identification of conditional ATEs in the target populati…On Causal Inference with Temporal and Spatial Spillovers in Panel Data¶
讲者: Anish Agarwal · 讨论人: Iavor Bojinov and Ashesh Rambachan · 2023-11-07
链接：视频 · 幻灯片 · arXiv
摘要
Panel data is a ubiquitous setting where one collects multiple measurements over time of a collection of heterogeneous units (e.g., individuals, firms, geographic entities). Two pervasive challenges in doing causal inference in such settings are: (1) temporal spillovers - interventions in the past affect current outcomes; (2) spatial spillovers - interventions a unit receives affects the outcomes of its neighboring units. We develop a causal framework to tackle these two challenges. We do so by …Confounder selection via iterative graph expansion¶
讲者: Richard Guo · 讨论人: Ilya Shpitser · 2023-10-31
链接：视频 · 幻灯片 · arXiv
摘要
Confounder selection, namely choosing a set of covariates to control for confounding between a treatment and an outcome, is arguably the most important step in the design of observational studies. Previous methods, such as Pearl's celebrated back-door criterion, typically require pre-specifying a causal graph, which can often be difficult in practice. We propose an interactive procedure for confounder selection that does not require pre-specifying the graph or the set of observed variables. This…Intervention Generalization: A View from Factor Graph Models¶
讲者: Ricardo Silva · 讨论人: Anish Agarwal · 2023-10-17
链接：视频 · 幻灯片 · arXiv
摘要
One of the goals of causal inference is to generalize from past experiments and observational data to novel conditions. While it is in principle possible to eventually learn a mapping from a novel experimental condition to an outcome of interest, provided a sufficient variety of experiments is available in the training data, coping with a large combinatorial space of possible interventions is hard. Under a typical sparse experimental design, this mapping is ill-posed without relying on heavy reg…How to learn more from observational factorial studies¶
讲者: Ruoqi Yu, Peng Ding · 讨论人: José Zubizarreta and Luke Keele [new format] · 2023-10-10
链接：视频
摘要
Many scientific questions in biomedical research, environmental sciences, and psychology involve understanding the impact of multiple factors on an outcome of interest. Randomized factorial experiments are a popular tool for evaluating the causal effects of multiple factors and their interactions simultaneously. However, randomization is often infeasible in many empirical studies, and drawing reliable causal inferences for multiple factors in observational studies remains challenging. As the num…Two fundamental problems in causal mediation analysis Discussant [standard form]: James Robins (Harvard University) and Thomas Richardson (University of Washington)¶
讲者: Caleb Miles · 2023-10-03
链接：视频 · 幻灯片
Two fundamental problems in causal mediation analysis （暂无精读）¶
讲者: Caleb Miles, - Discussant [standard form]:, James Robins, Thomas Richardson · 2023-10-03
链接：视频 · 幻灯片
摘要
Scientists are often interested in understanding mediating mechanisms that can help explain causal effects. A vast body of literature on mediation analysis has accumulated since two foundational articles (Robins and Greenland, 1992; Pearl, 2001) formalized mediation in the language of causality. However, causal mediation analysis poses many fundamental, interesting, and unresolved difficulties in its causal interpretation, nonparametric identification assumptions (which are much stronger than mo…Efficient Nonparametric Estimation of Stochastic Policy Effects with Clustered Interference¶
讲者: Michael Hudgens, Chanhwa Lee · 2023-09-26
链接：视频 · 幻灯片 · arXiv
摘要
Interference occurs when a unit's treatment (or exposure) affects another unit's outcome. In some settings, units may be grouped into clusters such that it is reasonable to assume that interference, if present, only occurs between individuals in the same cluster, i.e., there is clustered interference. Various causal estimands have been proposed to quantify treatment effects under clustered interference from observational data, but these estimands either entail treatment policies lacking real-wor…Better Than Difference in Differences¶
讲者: Andrew Gelman · 2023-09-19
链接：视频 · 幻灯片
摘要
It is not always clear how to adjust for control data in causal inference, balancing the goals of reducing bias and variance. We show how, in a setting with repeated experiments, Bayesian hierarchical modeling yields an adaptive procedure that uses the data to determine how much adjustment to perform. The result is a novel analysis with increased statistical efficiency compared with the default analysis based on difference estimates. The increased efficiency can have real-world consequences in t…Maintained by 陈星宇 · Homepage · Source