Causality and Generalizability: Identifiability and Learning Methods

Abstract

This PhD thesis contains several contributions to the field of statistical causal modeling. Statistical causal models are statistical models embedded with causal assumptions that allow for the inference and reasoning about the behavior of stochastic systems affected by external manipulation (interventions). This thesis contributes to the research areas concerning the estimation of causal effects, causal structure learning, and distributionally robust (out-of-distribution generalizing) prediction methods. We present novel and consistent linear and non-linear causal effects estimators in instrumental variable settings that employ data-dependent mean squared prediction error regularization. Our proposed estimators show, in certain settings, mean squared error improvements compared to both canonical and state-of-the-art estimators. We show that recent research on distributionally robust prediction methods has connections to well-studied estimators from econometrics. This connection leads us to prove that general K-class estimators possess distributional robustness properties. We, furthermore, propose a general framework for distributional robustness with respect to intervention-induced distributions. In this framework, we derive sufficient conditions for the identifiability of distributionally robust prediction methods and present impossibility results that show the necessity of several of these conditions. We present a new structure learning method applicable in additive noise models with directed trees as causal graphs. We prove consistency in a vanishing identifiability setup and provide a method for testing substructure hypotheses with asymptotic family-wise error control that remains valid post-selection. Finally, we present heuristic ideas for learning summary graphs of nonlinear time-series models.

arXiv commentPhD Thesis

References in corpus (28)

• Domain Adaptation for Statistical Classifiers
• Estimating high-dimensional directed acyclic graphs with the PC-algorithm
• Domain Generalization via Invariant Feature Representation
• Robust Physical-World Attacks on Deep Learning Models
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• On the Identifiability of the Post-Nonlinear Causal Model
• Assumptions of IV Methods for Observational Epidemiology
• Identifiability of Gaussian structural equation models with equal error variances
• DAGs with NO TEARS: Continuous Optimization for Structure Learning
• Kernel Granger causality and the analysis of dynamical networks
• Causal Inference in the Presence of Latent Variables and Selection Bias
• Gaussian Processes and Kernel Methods: A Review on Connections and Equivalences
• Distributionally Robust Logistic Regression
• Geometry of the faithfulness assumption in causal inference
• Invariant Models for Causal Transfer Learning
• Causal Inference on Discrete Data using Additive Noise Models
• High-dimensional instrumental variables regression and confidence sets
• Expanding the Transfer Entropy to Identify Information Subgraphs in Complex Systems
• Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributions
• Strong Faithfulness and Uniform Consistency in Causal Inference
• Learning Sparse Causal Models is not NP-hard
• A lower bound on the differential entropy of log-concave random vectors with applications
• Identifying confounders using additive noise models
• Learning stable and predictive structures in kinetic systems: Benefits of a causal approach
• From Ordinary Differential Equations to Structural Causal Models: the deterministic case
• Markov equivalence of marginalized local independence graphs
• Causal Discovery from Subsampled Time Series Data by Constraint Optimization
• Structural causal models for macro-variables in time-series