Fairness through Causal Awareness: Learning Causal Latent-Variable Models for Biased Data

By David Madras, Elliot Creager, Toniann Pitassi, Richard S. Zemel. 2018

146 citations recorded by the source corpus.

How do we learn from biased data? Historical datasets often reflect historical prejudices; sensitive or protected attributes may affect the observed treatments and outcomes. Classification algorithms tasked with predicting outcomes accurately from these datasets tend to replicate these biases. We advocate a causal modeling approach to learning from biased data, exploring the relationship between fair classification and intervention. We propose a causal model in which the sensitive attribute confounds both the treatment and the outcome. Building on prior work in deep learning and generative modeling, we describe how to learn the parameters of this causal model from observational data alone, even in the presence of unobserved confounders. We show experimentally that fairness-aware causal modeling provides better estimates of the causal effects between the sensitive attribute, the treatment, and the outcome. We further present evidence that estimating these causal effects can help learn policies that are both more accurate and fair, when presented with a historically biased dataset.

Subfields: Artificial Intelligence, Economics and Econometrics, Statistics and Probability, 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 Economics and Econometrics

In Statistics and Probability

In Causal Inference & Discovery

In AI Ethics, Fairness & Society

Citations in the corpus

Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias (2020, background): “…data (e.g., Lu et al., 2018; Hall Maudslay et al., 2019; Zhao et al., 2018a; Kaushik et al., 2019), the training process (e.g., Huang et al., 2019; Qian et al., 2019), or the model itself (e.g., Madras et al., 2019; Romanov et al., 2019; Gehrmann et al., 2019) to ensure counterfactual fairness.”
Causal Interpretability for Machine Learning - Problems, Methods and Evaluation (2020, background): “2], [8], [112], [81], [7], Example-based: [35], [45], [39], [101], [76], [86], [73], [61], [60], [33], [34] Post-hoc: [47], [89], [66], [93], [91], [108], [20], [16], [96], [62] Fairness: [53], [46], [67], [109] Guarantee: [49], [17], [88] Table 2: A summary of the state-of-the-art frameworks for ea”
Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence (2020, background): “…in response, have proposed fairness metrics that attempt to use causality (Chiappa and Isaac 2019; Mitchell et al. 2018; Nabi and Shpitser 2018; Madras et al. 2019) or interactivity (Canetti et al. 2019; Jung et al. 2019) to integrate more contextual awareness of human conceptions of fairness.”
Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection (2020, background): “Fairness in classiﬁcation can be deﬁned in many ways (Hardt et al., 2016; Madras et al., 2019; Zhang et al., 2018).”
Participatory Problem Formulation for Fairer Machine Learning Through Community Based System Dynamics (2020, methodology): “…Structural Causal Models (SCMs) and Causal Bayesian Networks (CBNs) that have been recently proposed to model causality in ML problem formulation (Madras et al., 2019; Chiappa & Isaac, 2018), CBSD is founded upon a system dynamics (SD) (Hovmand, 2014a) approach, which takes a characteristically…”
Awareness in practice: tensions in access to sensitive attribute data for antidiscrimination (2019, background): “However, when emphasizing the importance of awareness of sensitive attributes in developing and implementing fairnessenhancing interventions, fair ML research and toolkits [9, 38, 67] often take for granted when framing problems and their solutions that sensitive attribute data are available as inpu”
A survey on datasets for fairness‐aware machine learning (2021, background): “The reason is the causal eﬀects (Madras, Creager, Pitassi, & Zemel, 2019) of such attributes, including observable “proxy” attributes.”
Extending the Machine Learning Abstraction Boundary: A Complex Systems Approach to Incorporate Societal Context (2020, methodology): “ to be made. Often times these causal diagrams are constructed using directed acyclic graphs (DAGs). A number of these approaches leverage Structural Causal Models (SCMs) and Causal Bayesian Networks [18, 62] for measuring unfairness in datasets or building fair decision making models that have bias”
HARK Side of Deep Learning - From Grad Student Descent to Automated Machine Learning (2019, background): “Therefore, it is crucial to develop models that are fair and unbiased regardless of the biases in the data [59, 60].”
Investigating Gender Bias in Language Models Using Causal Mediation Analysis (2020, background): “…data (e.g., Lu et al., 2018; Hall Maud-slay et al., 2019; Zhao et al., 2018a; Kaushik et al., 2019), the training process (e.g., Huang et al., 2019; Qian et al., 2019), or the model itself (e.g., Madras et al., 2019; Romanov et al., 2019; Gehrmann et al., 2019) to ensure counterfactual fairness.”
Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems (2019, background): “Much of the recent progress in developing such decision making systems attempt to (approximately) learn in the presence of hidden confounders (Louizos et al., 2017; Madras et al., 2018) by estimating these confounders.”
Decolonizing Content Moderation: Does Uniform Global Community Standard Resemble Utopian Equality or Western Power Hegemony? (2023, background): “Second, the platforms need to adopt fairness-aware computing approaches to learn how historical biases shape moderation outcomes [89] and enact participatory design practices by allowing stakeholders, who are directly impacted by moderation outcomes, to define fairness in their own terms [73].”