A Causal Framework for Discovering and Removing Direct and Indirect Discrimination

By Lu Zhang, Yongkai Wu, Xintao Wu. 2016

191 citations recorded by the source corpus.

In this paper, we investigate the problem of discovering both direct and indirect discrimination from the historical data, and removing the discriminatory effects before the data is used for predictive analysis (e.g., building classifiers). The main drawback of existing methods is that they cannot distinguish the part of influence that is really caused by discrimination from all correlated influences. In our approach, we make use of the causal network to capture the causal structure of the data. Then we model direct and indirect discrimination as the path-specific effects, which accurately identify the two types of discrimination as the causal effects transmitted along different paths in the network. Based on that, we propose an effective algorithm for discovering direct and indirect discrimination, as well as an algorithm for precisely removing both types of discrimination while retaining good data utility. Experiments using the real dataset show the effectiveness of our approaches.

Subfields: Artificial Intelligence, Causal Inference & Discovery, AI Ethics, Fairness & Society

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 Causal Inference & Discovery

In AI Ethics, Fairness & Society

Citations in the corpus

A Survey on Bias and Fairness in Machine Learning (2019, methodology): “For instance, in [135] authors propose a causal-based framework that detects direct and indirect discrimination in the data along with their removal techniques.”
Attributing Fair Decisions with Attention Interventions (2021, background): “This phenomenon is called indirect discrimination (Zliobaite 2015; Hajian and Domingo-Ferrer 2013; Zhang, Wu, and Wu 2017).”
Learning Optimal Fair Policies (2018, background): “Though there has been growing interest in the issues of fairness in machine learning (Pedreshi et al., 2008; Feldman et al., 2015; Hardt et al., 2016; Kamiran et al., 2013; Corbett-Davies et al., 2017; Jabbari et al., 2017; Kusner et al., 2017; Zhang & Bareinboim, 2018; Mitchell & Shalden, 2018; Zha”
FairGAN: Fairness-aware Generative Adversarial Networks (2018, methodology): “Some methods mainly modify the labels of the dataset [13, 26].”
Path-Specific Counterfactual Fairness (2018, methodology): “It was recently argued (Qureshi et al., 2016; Bonchi et al., 2017; Kilbertus et al., 2017; Kusner et al., 2017; Nabi & Shpitser, 2018; Russell et al., 2017; Zhang et al., 2017; Zhang & Wu, 2017; Zhang & Bareinboim, 2018) that using a causal framework (Pearl, 2000; Dawid, 2007; Pearl et al.”
Fair Inference on Outcomes (2017, methodology): “We consider the more general case of bounding the PSE by ǫ l , ǫ u to link with earlier work in (Zhang, Wu, and Wu 2017), and for mathematical convenience.”
Achieving Non-Discrimination in Data Release (2016, background): “1 INTRODUCTION Discrimination discovery and prevention/removal has been an active research area recently [9, 12, 16, 28, 29, 39].”
Inclusive GAN: Improving Data and Minority Coverage in Generative Models (2020, background): “Bias in machine learning results from data imbalance, which can be detected and alleviated by three categories of approaches: The pre-process approaches that purify data from bias before training [7,12,13,51], the in-process approaches that enforce fairness during training with constraints or regula”
Achieving non-discrimination in prediction (2017, methodology): “Many methods have been proposed for constructing discrimination-free classifiers, which can be broadly classified into three categories: the pre-process methods that modify the training data [Kamiran and Calders, 2009a; Feldman et al., 2015; Zhang et al., 2017; Calmon et al., 2017; Nabi and Shpitser”
PC-Fairness: A Unified Framework for Measuring Causality-based Fairness (2019, methodology): “[19, 7, 16] O = ∅ or { S } and π = π i ⊂ Π Individual direct discrimination In the context of fair machine learning, we use S ∈ { s + , s − } to denote the protected attribute, Y ∈ { y + , y + } to denote the decision, and X to denote a set of non-protected attributes.”
Agent Incentives: A Causal Perspective (2021, background): “An incentive criterion corresponding to path-specific effects (Zhang, Wu, and Wu 2017; Nabi and Shpitser 2018) is deferred to future work.”
On conditional parity as a notion of non-discrimination in machine learning (2017, background): “Zhang, Wu and Wu (2016) and Nabi and Shpitser (2017) formalize discrimination as the presence of path speciﬁc eﬀects ( cf. Pearl (2009), § 4.5.3).”
Anti-discrimination learning: a causal modeling-based framework (2017, methodology): “In the discrimination removal algorithm in [42], it simply cuts off several paths in π i to dissatisfy the recanting witness criterion.”
Fairness in Machine Learning (2020, methodology): “…21, 22, 23, 25, 28, 29, 31, 34, 44, 45, 46, 52, 53, 54, 55, 56, 59, 66, 69, 71, 72, 73, 74, 75, 82, 84, 87, 89, 90, 94, 95, 97, 98, 99, 103, 104, 106, 114, 115, 116, 121, 123, 126, 138, 139, 141, 148, 149, 153, 156, 159, 165, 174, 176, 177, 178, 179, 180, 182, 183, 184, 186, 187, 191, 192, 193].”