Fall 2025 complete list with abstracts

See our homepage for future talks.

• Tuesday, Dec 09, 2025: Zoom link (webinar ID: 968 8371 7451, password: 414559)  
  - Speaker: Carlos Cinelli (University of Washington)
  - Title: Long Story Short: Omitted Variable Bias in Causal Machine Learning
  - Abstract: We develop a general theory of omitted variable bias for a wide range of common causal parameters, including (but not limited to) averages of potential outcomes, average treatment effects, average causal derivatives, and policy effects from covariate shifts. Our theory applies to nonparametric models, while naturally allowing for (semi-)parametric restrictions (such as partial linearity) when such assumptions are made. We show how simple plausibility judgments on the maximum explanatory power of omitted variables are sufficient to bound the magnitude of the bias, thus facilitating sensitivity analysis in otherwise complex, nonlinear models. Finally, we provide flexible and efficient statistical inference methods, which can leverage modern machine learning algorithms for estimation. These results allow empirical researchers to perform sensitivity analyses in a flexible class of machine-learned causal models using very simple, and interpretable, tools. Empirical examples demonstrate the utility of our approach.
  - Discussant: Dominik Rothenhäusler (Stanford University)
  [Slides][Paper][Video]

• Tuesday, Dec 02, 2025: OCIS+INI joint webinar (Zoom Link, webinar ID: 819 2387 7168​, password: Newton1).
  - Speaker: Rajen Shah (University of Cambridge)
  - Title: Hunt and test for assessing the fit of semiparametric regression models
  - Abstract: We consider testing the goodness of fit of semiparametric regression models, such as generalised additive models, partially linear models, and quantile additive regression models: a class of problems that includes, for example, testing for heterogeneous treatment effects. We propose an approach that involves splitting the data in two parts. On one part, we "hunt" for any signal that may be present in the score-type residuals following a fit of the null model. On the remaining data, we test for the presence of the potential signal thus found.  In the first, hunting stage of the procedure, our framework allows the use of any flexible regression method chosen by the practitioner, such as a random forest. The method can therefore harness the power of modern machine learning tools to detect complex alternatives. A challenge in the testing step is that any first-order bias in the residuals may lead to rejection under the null. To address this, we employ a debiasing step, which we show is equivalent to performing a particular weighted least squares regression. We establish that the type I error can be controlled under relatively mild conditions and that the test has power against alternatives for which, with high probability, the hunted signal is correlated with the true signal in the score residuals.
  - Discussant: Mats Stensrud (EPFL)
  [Slides][Video][Discussion slides]

• Tuesday, Nov 18, 2025: Caleb Miles (Columbia University)
  - Title: Addressing an extreme positivity violation to distinguish the causal effects of surgery and anesthesia via separable effects
  - Abstract: The U.S. Food and Drug Administration has cautioned that prenatal exposure to anesthetic drugs during the third trimester may have neurotoxic effects; however, there is limited clinical evidence available to substantiate this recommendation. One major scientific question of interest is whether such neurotoxic effects might be due to surgery, anesthesia, or both. Isolating the effects of these two exposures is challenging because they are observationally equivalent, thereby inducing an extreme positivity violation. To address this, we adopt the separable effects framework of Robins and Richardson (2010) to identify the effect of anesthesia (alone) by blocking effects through variables that are assumed to completely mediate the causal pathway from surgery to the outcome. We apply this approach to data from the nationwide Medicaid Analytic eXtract (MAX) from 1999 through 2013, which linked 16,778,281 deliveries to mothers enrolled in Medicaid during pregnancy. Furthermore, we assess the sensitivity of our results to violations of our key identification assumptions.
  - Discussant: James Robins (Harvard University) and Thomas Richardson (University of Washington)
  [Slides][Paper][Video][Discussion slides]

• Tuesday, Nov 11, 2025 [SCI+OCIS]:
  - Panel discussion: Funding Your Work: Thinking Beyond the NIH
  - Panelists: Michael Thompson (University of Michigan), Shu Yang (North Carolina State University)
  [Video]

• Tuesday, Nov 04, 2025: Zijun Gao (USC Marshall Business School)
  - Title: Explainability and Analysis of Variance
  - Abstract:  Existing tools for explaining complex models and systems are associational rather than causal and do not provide mechanistic understanding. We propose a new notion called counterfactual explainability for causal attribution that is motivated by the concept of genetic heritability in twin studies. Counterfactual explainability extends methods for global sensitivity analysis (including the functional analysis of variance and Sobol's indices), which assumes independent explanatory variables, to dependent explanations by using a directed acyclic graphs to describe their causal relationship. Therefore, this explanability measure directly incorporates causal mechanisms by construction. Under a comonotonicity assumption, we discuss methods for estimating counterfactual explainability and apply them to a real dataset dataset to explain income inequality by gender, race, and educational attainment.
  - Discussant: Art Owen (Stanford University)
  [Slides][Paper][Video][Discussion slides]

• Tuesday, Oct 28, 2025: Linbo Wang (University of Toronto)
  - Title: The synthetic instrument: From sparse association to sparse causation
  - Abstract:  In many observational studies, researchers are often interested in studying the effects of multiple exposures on a single outcome. Standard approaches for high-dimensional data such as the lasso assume the associations between the exposures and the outcome are sparse. These methods, however, do not estimate the causal effects in the presence of unmeasured confounding. In this paper, we consider an alternative approach that assumes the causal effects in view are sparse. We show that with sparse causation, the causal effects are identifiable even with unmeasured confounding. At the core of our proposal is a novel device, called the synthetic instrument, that in contrast to standard instrumental variables, can be constructed using the observed exposures directly. We show that under linear structural equation models, the problem of causal effect estimation can be formulated as an $\ell_0$-penalization problem, and hence can be solved efficiently using off-the-shelf software. Simulations show that our approach outperforms state-of-art methods in both low-dimensional and high-dimensional settings. We further illustrate our method using a mouse obesity dataset.
  - Discussant: Zijian Guo (Zhejiang University)
  [Slides][Paper][Video][Discussion slides]

• Tuesday, Oct 21, 2025: Wooseok Ha (Korea Advanced Institute of Science and Technology)
  - Title: When few labeled target data suffice: a theory of semi-supervised domain adaptation via fine-tuning from multiple adaptive starts
  - Abstract:   Semi-supervised domain adaptation (SSDA) aims to achieve high predictive performance in the target domain with limited labeled target data by exploiting abundant source and unlabeled target data. Despite its significance in numerous applications, theory on the effectiveness of SSDA remains largely unexplored, particularly in scenarios involving various types of source-target distributional shifts. In this talk, I will present a theoretical framework based on structural causal models (SCMs) which allows us to analyze and quantify the performance of SSDA methods when labeled target data is limited. Within this framework, I introduce three SSDA methods, each having a fine-tuning strategy tailored to a distinct assumption about the source and target relationship. Under each assumption, I demonstrate how extending an unsupervised domain adaptation (UDA) method to SSDA can achieve minimax-optimal target performance with limited target labels. Finally, when the relationship between source and target data is only vaguely known—a common practical concern—I will describe the Multi Adaptive-Start FineTuning (MASFT) algorithm, which fine-tunes UDA models from multiple starting points and selects the best-performing one based on a small hold-out target validation dataset. Combined with model selection guarantees, MASFT achieves near-optimal target predictive performance across a broad range of types of distributional shifts while significantly reducing the need for labeled target data.
  - Discussant: Jason Kluswoski (Princeton University)
  [Paper][Slides][Video][Discussion slides]

• Tuesday, Oct 14, 2025: Jakob Runge (University of Potsdam)
  - Title: Causal Inference on Time Series Data with the Tigramite Package
  - Abstract: This talk introduces the open-source Python package Tigramite, which implements constraint-based algorithms such as PCMCI+ and many variants thereof as methods optimised for causal discovery on time series. In addition, Tigramite features causal effect estimation using optimal adjustment. I will outline the basic ideas behind PCMCI and optimal adjustment and then demonstrate practical workflows in Tigramite, including a user-friendly guide to choosing methods in causal inference based on causal questions, assumptions and available data. I look forward to feedback and exchange on improvements of the package to make causal inference accessible for practitioners dealing with time series data.
  [Slides][Package][Video]

• Tuesday, Oct 07, 2025  [SCI+OCIS]: Miguel Hernán (Harvard University)
  - Title: Making decisions is hard but making decisions without data is much harder: How causal inference research helped governments during the last pandemic
  - Abstract: The first question a decision maker asks is “Do we have a problem?”; the second one is “How do we handle the problem?”. Answering the first question requires descriptive studies; answering the second one requires causal studies. This talk describes examples of how this process worked in the real world during the last pandemic. It is partly based on my experience as an embedded researcher in a government agency. Some take-home messages are: conducting good descriptive studies is difficult but indispensable; actionable causal inference can sometimes rely on randomized trials but will often have to rely on observational emulations of trials; sometimes the causal questions are so complex that only mathematical models will help decision makers; and researchers are usually not qualified to tell decision makers which decisions they should make.
  [Slides]

• Tuesday, Sep 30, 2025: Joseph Antonelli (University of Florida)
  - Title: Partial identification and unmeasured confounding with multiple treatments and multiple outcomes
  - Abstract: Estimating the health effects of multiple air pollutants is a crucial problem in public health, but one that is difficult due to unmeasured confounding bias. Motivated by this issue, we develop a framework for partial identification of causal effects in the presence of unmeasured confounding in settings with multiple treatments and multiple outcomes. Under a factor confounding assumption, we show that joint partial identification regions for multiple estimands can be more informative than considering partial identification for individual estimands one at a time. We show how assumptions related to the strength of confounding or magnitude of plausible effect sizes for one estimand can reduce the partial identification regions for other estimands. As a special case of this result, we explore how negative control assumptions reduce partial identification regions and discuss conditions under which point identification can be obtained. We develop novel computational approaches to finding partial identification regions under a variety of these assumptions. We then estimate the causal effect of PM2.5 components on a variety of public health outcomes in the United States Medicare cohort, where we find that, in particular, the detrimental effect of black carbon is robust to the potential presence of unmeasured confounding bias.
  - Discussant: Carlos Cinelli (University of Washington)
  [Slides][Paper][Discussant slides][Video]

• Tuesday, Sep 23, 2025: Young researchers' seminar
  Speaker 1: Jiaqi Zhang (MIT)
  - Title: Learning causal cellular programs from large-scale perturbations
  - Abstract: Complex molecular mechanisms govern cellular functions in living organisms and shape their behavior in health and disease. Understanding these mechanisms can greatly accelerate therapeutic discovery, yet it remains challenging due to the high dimensionality and intricate dependencies within biological systems. Recent advances in experimental technologies, however, are beginning to make this problem more tractable. For example, we can now systematically perturb individual or combinations of genes in single cells and measure their downstream effects, enabling empirical identification and validation of causal relationships. Nevertheless, perturbation data remain noisy and high-dimensional, with effects often sparse and subtle.
  In this talk, I will present our work addressing two key challenges in this emerging paradigm: (1) how to define and learn the causal programs that govern high-dimensional or perceptual data; (2) how to model perturbational effects in a way that enables the prediction of novel perturbations. For (1), we introduce new causal representation theories that guarantee identifiability of the underlying causal programs, given sufficient samples and regularizing condition. For (2), we develop a modular framework that models perturbation effects through distributional discrepancies. This approach captures nuanced, sample-level changes and enables extrapolation to predict the effects of unseen perturbations across diverse conditions and data types.
  [Slides][Papers: #1, #2][Video]
  Speaker 2: Sizhu Lu (UC Berkeley)
  - Title: Estimating treatment effects with competing intercurrent events in randomized controlled trials
  - Abstract: The analysis of randomized controlled trials is often complicated by intercurrent events (ICEs) -- events that occur after treatment initiation and affect either the interpretation or existence of outcome measurements. Examples include treatment discontinuation or the use of additional medications. In two recent clinical trials for systemic lupus erythematosus with complications of ICEs, we classify the ICEs into two broad categories: treatment-related (e.g., treatment discontinuation due to adverse events or lack of efficacy) and treatment-unrelated (e.g., treatment discontinuation due to external factors such as pandemics or relocation). To define a clinically meaningful estimand, we adopt tailored strategies for each category of ICEs. For treatment-related ICEs, which are often informative about a patient's outcome, we use the composite variable strategy that assigns an outcome value indicative of treatment failure. For treatment-unrelated ICEs, we apply the hypothetical strategy, assuming their timing is conditionally independent of the outcome given treatment and baseline covariates, and hypothesizing a scenario in which such events do not occur. A central yet previously overlooked challenge is the presence of competing ICEs, where the first ICE censors all subsequent ones. Despite its ubiquity in practice, this issue has not been explicitly recognized or addressed in previous data analyses due to the lack of rigorous statistical methodology. In this paper, we propose a principled framework to formulate the estimand, establish its nonparametric identification and semiparametric estimation theory, and introduce weighting, outcome regression, and doubly robust estimators. We apply our methods to analyze the two systemic lupus erythematosus trials, demonstrating the robustness and practical utility of the proposed framework.
  [Slides][Paper][Video]