Fairness through awareness
By Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, Richard S. Zemel. 2011
4,555 citations recorded by the source corpus.
We study fairness in classification, where individuals are classified, e.g., admitted to a university, and the goal is to prevent discrimination against individuals based on their membership in some group, while maintaining utility for the classifier (the university). The main conceptual contribution of this paper is a framework for fair classification comprising (1) a (hypothetical) task-specific metric for determining the degree to which individuals are similar with respect to the classification task at hand; (2) an algorithm for maximizing utility subject to the fairness constraint, that similar individuals are treated similarly. We also present an adaptation of our approach to achieve the complementary goal of "fair affirmative action," which guarantees statistical parity (i.e., the demographics of the set of individuals receiving any classification are the same as the demographics of the underlying population), while treating similar individuals as similarly as possible. Finally, we discuss the relationship of fairness to privacy: when fairness implies privacy, and how tools developed in the context of differential privacy may be applied to fairness.
Subfields: Artificial Intelligence, Safety Research, Sociology and Political Science, 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
- Percy Liang
- Yoshua Bengio
- Mohit Bansal
- Quoc V. Le
- Sergey Levine
- Jianfeng Gao
- Ruslan Salakhutdinov
- Luke S. Zettlemoyer
In Safety Research
In Sociology and Political Science
- Silvia Chiappa
- Krishna P. Gummadi
- Xintao Wu
- Lu Zhang
- Aaron Roth
- Indrė Žliobaitė
- Luca Oneto
- Muhammad Bilal Zafar
In Safety, Robustness & Interpretability
- Silvia Chiappa
- Dana Pessach
- Erez Shmueli
- Percy Liang
- Adrian Weller
- Nicolas Papernot
- Nicholas Carlini
- Krishna P. Gummadi
Citations in the corpus
- Reducing Sentiment Bias in Language Models via Counterfactual Evaluation (2019, methodology): “3.1 in Dwork et al. (2012)).”
- Societal Biases in Language Generation: Progress and Challenges (2021, background): “Language generation tasks often involve stochastic generation of open-ended and lengthy texts, traits that are not directly compatible with traditional algorithmic bias definitions (e.g., equalized odds, equal opportunity, demographic parity (Dwork et al., 2012; Hardt et al., 2016)).”
- Social Biases in NLP Models as Barriers for Persons with Disabilities (2020, methodology): “Following Dwork et al. (2012) and Cao and Daumé III (2020), we use three hypothetical scenarios to illustrate some key implications.”
- Shortcut learning in deep neural networks (2020, background): “ful approach for learning policies that generalise to similar situations in the real-world [70]. 15 Fairness Fairness research aims at making machine decisions “fair” according to a certain definition [131]. Individual fairness aims at treating similar individuals similarly while group fairness aims ”
- Measurement and Fairness (2019, background): “g., with similar false positive rates) to another group. Statistically, there is necessarily incompatibility between optimizing for individual outcomes vs. group-level outcomes [13, 12]. Dwork et al. [18] argue that harms emerge for the individual using traditional group-based operationalizations of”
- On The Impact of Machine Learning Randomness on Group Fairness (2023, background): “Fair machine learning can be broadly divided into two categories, (i) group fairness [15], and (ii) individual fairness [19].”
- Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (2016, background): “Approaches to modify classification algorithms to define and achieve various notions of fairness have been described in a number of works, see, e.g., [3, 7, 10] and a recent survey [41].”
- Towards the Science of Security and Privacy in Machine Learning (2016, methodology): “To learn fair models, Zemel et al. first learn an intermediate representation that encodes a sanitized variant of the data, as first discussed in [109].”
- Does Object Recognition Work for Everyone? (2019, background): “The demographic parity formulation of fairness does not focus on group membership, but are based on the idea that similar individuals should receive similar predictions [12].”
- Women also Snowboard: Overcoming Bias in Captioning Models (2018, background): “, age, gender, sexual orientation) in a fair manner is increasingly important as AI continues to become more ubiquitous [14,15,16,17].”
- Towards fairer datasets: filtering and balancing the distribution of the people subtree in the ImageNet hierarchy (2019, background): “Researchers have proposed multiple fairness metrics including statistical parity [12, 13, 24, 39], disparate impact [27, 80], equalized odds [34] and individual fairness [23], and analyzed the relationship between them [43, 60].”
- Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints (2017, background): “Various approaches have been proposed to make a “fair” binary classification (Baro-cas and Selbst, 2014; Dwork et al., 2012; Feldman et al., 2015; Zliobaite, 2015).”
- Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias (2020, methodology): “Another extension of this work for training a fair model would be to use the individual fairness framework (Dwork et al., 2011) that inspired some of the alternate metrics discussed in this article.”
- Counterfactual Fairness in Text Classification through Robustness (2018, methodology): “r of downsides, including predicting using unnatural, semantically incorrect inputs, losing information to contextualize inputs related to sensitive groups, discriminating based on redundant encoding [3], and difficulty being applied to more complex counterfactual settings. Sahaj Garg, Vincent Perot”
- Equality of Opportunity in Supervised Learning (2016, background): “Unfortunately, as was already argued by Dwork et al. [DHP + 12], the notion is seriously flawed on two counts.”
- Towards A Rigorous Science of Interpretable Machine Learning (2017, background): “Some areas, such as the fairness [Hardt et al., 2016] and privacy [Toubiana et al., 2010, Dwork et al., 2012, Hardt and Talwar, 2010] the research communities have formalized their criteria, and these formalizations have allowed for a blossoming of rigorous research in these fields (without the need…”