Avoiding Discrimination through Causal Reasoning

By Niki Kilbertus, Mateo Rojas-Carulla, Giambattista Parascandolo, Moritz Hardt, Dominik Janzing, Bernhard Schölkopf. 2017

640 citations recorded by the source corpus.

Recent work on fairness in machine learning has focused on various statistical discrimination criteria and how they trade off. Most of these criteria are observational: They depend only on the joint distribution of predictor, protected attribute, features, and outcome. While convenient to work with, observational criteria have severe inherent limitations that prevent them from resolving matters of fairness conclusively. Going beyond observational criteria, we frame the problem of discrimination based on protected attributes in the language of causal reasoning. This viewpoint shifts attention from "What is the right fairness criterion?" to "What do we want to assume about the causal data generating process?" Through the lens of causality, we make several contributions. First, we crisply articulate why and when observational criteria fail, thus formalizing what was before a matter of opinion. Second, our approach exposes previously ignored subtleties and why they are fundamental to the problem. Finally, we put forward natural causal non-discrimination criteria and develop algorithms that satisfy them.

Subfields: Artificial Intelligence, Safety Research, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety Research

In Safety, Robustness & Interpretability

Citations in the corpus

Reducing Sentiment Bias in Language Models via Counterfactual Evaluation (2019, background): “, 2012), counterfactual fairness requires the same model predictions before and after intervention on sensitive attributes in data-generating causal graphs (Kusner et al., 2017; Kilbertus et al., 2017; Chiappa, 2019; Chiappa and Isaac, 2019).”
Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems (2018, background): “Approaches on how to mitigate inappropriate biases (Schmidt, 2015; Bolukbasi et al., 2016; Kilbertus et al., 2017; Ryu et al., 2017; Speer, 2017; Zhang et al., 2018; Zhao et al., 2018) are also beyond the scope of this paper.”
Counterfactual Fairness in Text Classification through Robustness (2018, background): “ncorporate any external understanding about what is causally acceptable in making predictions. The limitations of observational criteria can be addressed by counterfactual or individual fairness, see [3, 9, 12]. Social science literature on fairness raises arguments for counterfactual reasoning as w”
Mitigating Gender Bias in Natural Language Processing: Literature Review (2019, background): “…to NLP; other ﬁelds in computer science such as data mining, machine learning, and security also study gender bias (Calders and Verwer, 2010; Feldman et al., 2015; Hardt et al., 2016; Misra et al., 2016; Kleinberg et al., 2016; Pleiss et al., 2017; Beutel et al., 2017; Kilbertus et al., 2017).”
Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Setting (2019, background): “” Although the latter question is arguably more interesting, it cannot be answered without additionally changing all other factors that are correlated with gender [27].”
A Survey on Bias and Fairness in Machine Learning (2019, methodology): “PCA [111] Machine learning Community detection [85] Machine learning Clustering [30] [7] Machine learning Graph embedding [21] Machine learning Causal inference [81] [136] [137] [132] [95] [94] [134] [69] [105] [133] Natural language processing Word embedding [19] [141] [50] [22] [138] Natural langu”
Counterfactual Fairness (2017, background): “When unfairness is judged to follow only some “pathways” in the causal graph (in a sense that can be made formal, see [21, 27]), nonparametric assumptions about the independence of counterfactuals may suffice, as discussed by [27].”
Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “Relatedly, No Unresolved Discrimination states that there should exist no directed path from A to D unless through a resolving variable (a variable that is influenced by A in a manner that we accept as nondiscriminatory) [70].”
Causal Modeling-Based Discrimination Discovery and Removal: Criteria, Bounds, and Algorithms (2019, background): “In addition, [24] suffers inherent limitations: (1) its proposed discrimination criteria can only qualitatively determine the existence of the discrimination, but”
Uncovering and Mitigating Algorithmic Bias through Learned Latent Structure (2019, background): “…history in machine learning, including Expectation-Maximization (Bailey, Elkan, and others 1994), topic modelling (Blei 2012), latent-SVM (Felzenszwalb, McAllester, and Ramanan 2008), and more recently, variational autoen-coders (VAE) (Kingma and Welling 2013; Rezende, Mo-hamed, and Wierstra 2014).”
Causal Reasoning for Algorithmic Fairness (2018, methodology): “A common theme of the alternatives to counterfactual fairness [21, 28, 44] is the focus on path-speciﬁc eﬀects , which was only brieﬂy mentioned in our formulation.”
Delayed Impact of Fair Machine Learning (2018, neutral): “This can be seen as relaxing the requirements compared with recent work on avoiding discrimination through causal reasoning that often required stronger assumptions [Kusner et al., 2017, Nabi and Shpitser, 2017, Kilbertus et al., 2017].”
Path-Specific Counterfactual Fairness (2018, methodology): “Kilbertus et al. (2017) and Nabi & Shpitser (2018) propose to constrain the learning of the model parameters such that the unfair effect is eliminated or reduced.”
On Fairness and Calibration (2017, background): “Fairness has also been considered in online learning [21, 24], unsupervised learning [5], and causal inference [25, 27].”
Towards Causal Representation Learning (2021, neutral): “include robust, explainable, or fair [142, 133, 276, 130, 260].”