Fairness and Machine Learning Limitations and Opportunities
By Solon Barocas, Moritz Hardt, Arvind Narayanan. 2018
Subfields: 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 Safety, Robustness & Interpretability
- Vinodkumar Prabhakaran
- Emily L. Denton
- Ben Hutchinson
- Sunipa Dev
- Sara Hooker
- Mark Díaz
- John Wieting
- Katherine Lee
Citations in the corpus
- Social Biases in NLP Models as Barriers for Persons with Disabilities (2020, background): “However, normative and social justifications should be considered when applying a statistical definition of fairness (Barocas et al., 2018; Blodgett et al., 2020).”
- PaLM 2 Technical Report (2023, background): “Harms are the adverse lived experiences of an algorithmic system’s deployment and use in the world (Barocas et al., 2017; Shelby et al., 2023), and downstream uses of a language model create risks of a range of potential harms (Blodgett et al.”
- Capabilities of GPT-4 on Medical Challenge Problems (2023, background): “However, engineers and organizations continue to face conceptual challenges with defining measures of fairness [JW21, BHN19].”
- Bias and Fairness in Large Language Models: A Survey (2023, methodology): “Here we attempt to disambiguate the types of harms that may emerge from LLMs, building on the definitions in machine learning works by Barocas et al. (2019), Bender et al.”
- Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models (2023, background): “It is essential to ensure that the training data used to develop AI models is representative of the diverse range of perspectives, experiences, and backgrounds that exist within society [96, 33, 90, 56, 97].”
- FAIR-Ensemble: When Fairness Naturally Emerges From Deep Ensembling (2023, background): “Beyond Top-line metrics Discussions of algorithmic bias often focus on datasets collection and curation (Barocas et al., 2019; Zhao et al., 2017; Shankar et al., 2017), with limited work to-date understanding the role of model design or optimization choices on amplifying or curbing bias (Ogueji et…”