OCIS · Fall 2024¶
- 共 8 场 · 8 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。
这一季的 8 场报告围绕三条主线展开：实验设计与复杂处理效应检验（Panos Toulis/Wenxuan Guo、Tianchen Qian、Toru Kitagawa）、异质性建模与加权视角（Jared S. Murray、Oliver Dukes、Yiqing Xu）、以及因果结构在域迁移与组合干预中的应用（Alexis Bellot、Anish Agarwal）。其中，加权视角与半参数效率的交叉、以及时间序列/动态设定下的因果推断是贯穿多场的隐性主题。
实验设计与复杂处理效应是最大的一条线。Panos Toulis/Wenxuan Guo 将 Fisherian 随机化检验与 ML 检验统计量结合，专门针对 A/B 实验中的异质性和溢出效应提升功效，是经典 FRT 在复杂场景下的现代化。Tianchen Qian 则聚焦移动健康中的微随机试验（MRT），用功能数据分析（FDA）建模时变因果效应，处理纵向高维历史过程的挑战。Toru Kitagawa 将政策学习（EWM）从横截面扩展到时间序列，提出 T-EWM 方法，直接处理动态最优策略选择。这三场共享“实验/准实验设计 + 非参数/ML 工具”的基因，但分别处理空间（溢出）、时间（时变）、和策略（动态决策）三个维度。
异质性建模与加权视角是另一条强线。Jared S. Murray 提供了一个统一视角：许多因果 ML 方法（核方法、高斯过程、BART、BCF）本质上等价于加权估计，权重对应 Riesz 表示，这为诊断模型失效和设计新方法提供了数学框架。Oliver Dukes 则从检验角度切入，提出非参数异质性检验，区分真实 CATE 变化与噪声，尤其关注政策制定者关心的定性异质性（方向相反）。Yiqing Xu 的 Factorial DID 看似是 DID 的变体，实则触及异质性的识别问题：当所有单元都暴露于事件时，标准 DID 只能识别效应修正而非因果调节，需要更苛刻的假设。这三场从“如何估计权重”、“如何检验异质性”、“如何识别因果调节”三个角度切入同一核心问题。
因果结构与域迁移/组合干预是第三大主线。Alexis Bellot 将可运输性理论从点识别推广到部分识别，处理域泛化中目标域分布不确定性的问题，用选择图编码跨域机制变化，给出性能保证的区间。Anish Agarwal 则处理组合干预的因果推断，利用低秩性（矩阵补全）和稀疏性（Fourier 基）的双重结构，在 N×2^P 个参数中实现高效学习。这两场都依赖因果图或潜在结果的结构假设（选择图、低秩/稀疏），但分别面向“预测迁移”和“干预效应估计”两个任务。
若想快速切入，建议按以下路径：基础入口——先看 Jared S. Murray（统一加权视角，是理解多场方法的数学基础）和 Panos Toulis/Wenxuan Guo（经典 FRT 的现代应用，实验设计入门）。进阶入口——若对动态/时间序列感兴趣，看 Toru Kitagawa（T-EWM）和 Tianchen Qian（MRT + FDA）；若对异质性检验与识别感兴趣，看 Oliver Dukes（非参数检验）和 Yiqing Xu（Factorial DID 的识别挑战）；若对结构迁移与组合干预感兴趣，看 Alexis Bellot（部分可运输性）和 Anish Agarwal（低秩+稀疏的组合干预）。
报告列表¶
ML-assisted Randomization Tests for Complex Treatment Effects in A/B Experiments¶
讲者: Panos Toulis, Wenxuan Guo · 讨论人: Xinran Li · 2024-12-10
链接：视频 · 幻灯片
摘要
Experimentation is widely used for causal inference and data-driven decision making across disciplines. In an A/B experiment, for example, a business randomizes two different treatments (e.g., website designs) to their customers and then aims to infer which treatment is better. In this paper, we construct randomization tests for complex treatment effects, including heterogeneity and interference. A key feature of our approach is the use of flexible machine learning (ML) models, where the ANOVA-l…Factorial Difference-in-Differences¶
讲者: Yiqing Xu · 讨论人: Erin Hartman · 2024-12-03
链接：视频 · 幻灯片 · arXiv
摘要
In many social science applications, researchers use the difference-in-differences (DID) estimator to establish causal relationships, exploiting cross-sectional variation in a baseline factor and temporal variation in exposure to an event that presumably may affect all units. This approach, which we term factorial DID (FDID), differs from canonical DID in that it lacks a clean control group unexposed to the event after the event occurs. In this paper, we clarify FDID as a research design in term…A Unifying Weighting Perspective on Causal Machine Learning: Kernel Methods, Gaussian Processes, and Bayesian Tree Models¶
讲者: Jared S. Murray · 讨论人: Rahul Singh · 2024-11-19
链接：视频
摘要
Causal machine learning methods based on kernel methods are powerful tools for estimating heterogeneous treatment effects; examples include kernel ridge regression, (causal) random forests, and many neural networks. A known but underappreciated result is that many of these methods have an equivalent representation as weighting estimators, with weights that correspond to an estimate of the Riesz representer of the estimand. This paper catalogs results about the weighting representation of heterog…Causal inference and machine learning in mobile health – modeling time-varying effects using longitudinal functional data¶
讲者: Tianchen Qian · 讨论人: Walter Dempsey · 2024-11-12
链接：视频 · 幻灯片 · arXiv
摘要
To optimize mobile health interventions and advance domain knowledge on intervention design, it is critical to understand how the intervention effect varies over time and with contextual information. This study aims to assess how a push notification suggesting physical activity influences individuals’ step counts using data from the HeartSteps micro-randomized trial (MRT). The statistical challenges include the time-varying treatments and longitudinal functional step count measurements. We propose…Policy Choice in Time-Series by Empirical Welfare Maximization¶
讲者: Toru Kitagawa · 讨论人: Mikkel Plagborg-Moller · 2024-10-29
链接：视频 · 幻灯片 · arXiv
摘要
This paper develops a novel method for policy choice in a dynamic setting where the available data is a multivariate time series. Building on the statistical treatment choice framework, we propose Time-series Empirical Welfare Maximization (T-EWM) methods to estimate an optimal policy rule for the current period or over multiple periods by maximizing an empirical welfare criterion constructed using nonparametric potential outcome time-series. We characterize conditions under which T-EWM consiste…Partial Transportability for Domain Generalization¶
讲者: Alexis Bellot · 讨论人: Adam Li · 2024-10-22
链接：视频 · 幻灯片
摘要
A fundamental task in AI is providing performance guarantees for predictions made in unseen domains. In practice, there can be substantial uncertainty about the distribution of new data, and corresponding variability in the performance of existing predictors. For example, a risk prediction tool fine-tuned on a patient population (e.g. particular hospital, geographic location) may not be equally optimal if deployed on a different patient population that may differ in several aspects. This talk st…Nonparametric tests of treatment effect homogeneity for policy-makers¶
讲者: Oliver Dukes · 讨论人: Edward Kennedy · 2024-10-15
链接：视频 · 幻灯片 · arXiv
摘要
Recent work has focused on nonparametric estimation of conditional treatment effects, but inference has remained relatively unexplored. We propose a class of nonparametric tests for both quantitative and qualitative treatment effect heterogeneity. The tests can incorporate a variety of structured assumptions on the conditional average treatment effect, allow for both continuous and discrete covariates, and do not require sample splitting. Furthermore, we show how the tests are tailored to detect…Synthetic Combinations: A Causal Inference Framework for Combinatorial Interventions¶
讲者: Anish Agarwal · 讨论人: Christina Lee Yu · 2024-10-01
链接：视频 · 幻灯片 · arXiv
摘要
We consider a setting where there are N heterogeneous units and p interventions. Our goal is to learn unit-specific potential outcomes for any combination of these p interventions, i.e., N×2^p causal parameters. Choosing a combination of interventions is a problem that naturally arises in a variety of applications such as factorial design experiments and recommendation engines (e.g., showing a set of movies that maximizes engagement for a given user). Running N×2^p experiments to estimate the va…Maintained by 陈星宇 · Homepage · Source