On Causal and Anticausal Learning

Abstract

We consider the problem of function estimation in the case where an underlying causal model can be inferred. This has implications for popular scenarios such as covariate shift, concept drift, transfer learning and semi-supervised learning. We argue that causal knowledge may facilitate some approaches for a given problem, and rule out others. In particular, we formulate a hypothesis for when semi-supervised learning can help, and corroborate it with empirical results.

Appears in Proceedings of the 29th International Conference on Machine Learning (ICML 2012). arXiv admin note: substantial text overlap with arXiv:1112.2738

References in corpus (1)

• Robust Learning via Cause-Effect Models

Cited by in corpus (70)

• Shortcut Learning in Deep Neural Networks
• Distinguishing cause from effect using observational data: methods and benchmarks
• Invariant Risk Minimization
• Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift
• Improving robustness against common corruptions by covariate shift adaptation
• Causal Discovery with Continuous Additive Noise Models
• Causality for Machine Learning
• Causal interpretation rules for encoding and decoding models in neuroimaging
• Detecting and Correcting for Label Shift with Black Box Predictors
• Weakly-Supervised Disentanglement Without Compromises
• A causal view of compositional zero-shot recognition
• Causal Intervention for Weakly-Supervised Semantic Segmentation
• Domain Adaptation with Conditional Distribution Matching and Generalized Label Shift
• Disentangling Factors of Variation Using Few Labels
• Learning Causal Semantic Representation for Out-of-Distribution Prediction
• Causality in cognitive neuroscience: concepts, challenges, and distributional robustness
• Towards a Learning Theory of Cause-Effect Inference
• Learning Neural Causal Models from Unknown Interventions
• Learning stable and predictive structures in kinetic systems: Benefits of a causal approach
• Invariant Causal Prediction for Block MDPs
• A Causal View on Robustness of Neural Networks
• CoPhy: Counterfactual Learning of Physical Dynamics
• Instance-dependent Label-noise Learning under a Structural Causal Model
• Invariant Risk Minimization Games
• Transfer Learning with Label Noise
• Are Generative Classifiers More Robust to Adversarial Attacks?
• Domain Adaptation as a Problem of Inference on Graphical Models
• A Prototype-Oriented Framework for Unsupervised Domain Adaptation
• Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual
• Can Subnetwork Structure be the Key to Out-of-Distribution Generalization?
• Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
• Domain adaptation under structural causal models
• Causal Discovery from Heterogeneous/Nonstationary Data with Independent Changes
• The Randomized Causation Coefficient
• Distinguishing Cause from Effect Based on Exogeneity
• Semi-Supervised Learning, Causality and the Conditional Cluster Assumption
• Semi-Generative Modelling: Covariate-Shift Adaptation with Cause and Effect Features
• Preventing Failures Due to Dataset Shift: Learning Predictive Models That Transport
• Detecting non-causal artifacts in multivariate linear regression models
• Robust Learning with the Hilbert-Schmidt Independence Criterion
• Online Adaptation to Label Distribution Shift
• Causal Discovery Using Proxy Variables
• Improvability Through Semi-Supervised Learning: A Survey of Theoretical Results
• Distributional robustness of K-class estimators and the PULSE
• Nonuniform Negative Sampling and Log Odds Correction with Rare Events Data
• Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Tests
• Deep causal representation learning for unsupervised domain adaptation
• Causality-aware counterfactual confounding adjustment for feature representations learned by deep models
• Adaptive Skip Intervals: Temporal Abstraction for Recurrent Dynamical Models
• Causality-aware counterfactual confounding adjustment as an alternative to linear residualization in anticausal prediction tasks based on linear learners
• Removing systematic errors for exoplanet search via latent causes
• Distinguishing Cause from Effect Using Quantiles: Bivariate Quantile Causal Discovery
• Linear Regression Games: Convergence Guarantees to Approximate Out-of-Distribution Solutions
• Target contrastive pessimistic risk for robust domain adaptation
• An Information-theoretic Approach to Distribution Shifts
• Nonparametric Causal Feature Selection for Spatiotemporal Risk Mapping of Malaria Incidence in Madagascar
• Active Learning under Label Shift
• Efficient Neural Causal Discovery without Acyclicity Constraints
• Optimization-based Causal Estimation from Heterogenous Environments
• Merging joint distributions via causal model classes with low VC dimension
• Data-driven causal path discovery without prior knowledge - a benchmark study
• Domain Generalization under Conditional and Label Shifts via Variational Bayesian Inference
• Predicting Counterfactuals from Large Historical Data and Small Randomized Trials
• A Label Proportions Estimation Technique for Adversarial Domain Adaptation in Text Classification
• Causal inference for data-driven debugging and decision making in cloud computing
• Causally Denoise Word Embeddings Using Half-Sibling Regression
• A Causal Direction Test for Heterogeneous Populations
• Disentangling Action Sequences: Discovering Correlated Samples
• Cycles in Causal Learning
• GalilAI: Out-of-Task Distribution Detection using Causal Active Experimentation for Safe Transfer RL