Bias and Fairness in NLP Models (4/22/2021)
Guest lecture by Divyansh Kaushik.
- Types of Bias in NLP Models
- How to Prevent Bias in NLP
- Fairness: Quantifying and Encouraging
References
- Reference: How Much Reading Does Reading Comprehension Require? A Critical Investigation of Popular Benchmarks (Kaushik et al. 2018)
- Reference: Adversarial Examples for Evaluating Reading Comprehension Systems (Jia and Liang 2017)
- Reference: Measuring and Mitigating Unintended Bias in Text Classification (Dixon et al. 2017)
- Reference: Counterfactual Thought (Byrne 2016)
- Reference: Norms in Counterfactual Selection (Fazelpour 2020)
- Reference: Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (Bolukbasi et al. 2016)
- Reference: Semantics derived automatically from language corpora contain human-like biases (Caliskan et al. 2017)
- Reference: On Measuring Social Biases in Sentence Encoders (May et al. 2019)
- Reference: Assessing Social and Intersectional Biases in Contextualized Word Representations (Tan and Celis 2019)
- Reference: Racial disparities in automated speech recognition (Koenecke et al. 2020)
- Reference: Learning Fair Representations (Zemel et al. 2013)
- Reference: Unsupervised Domain Adaptation by Backpropagation (Ganin and Lempitsky 2015)
- Reference: Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them (Gonen and Goldberg 2019)
- Reference: Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection (Ravgofel et al. 2020)
- Reference: Gender Bias in Neural Natural Language Processing (Lu et al. 2018)
- Reference: Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology (Zmigrod et al. 2018)
- Reference: Learning The Difference That Makes A Difference With Counterfactually-Augmented Data (Kaushik et al. 2020)
- Reference: Explaining the Efficacy of Counterfactually Augmented Data (Kaushik et al. 2021)
- Reference: Tie-breaker: Using language models to quantify gender bias in sports journalism (Fu et al. 2016)
- Reference: Unsupervised Discovery of Gendered Language through Latent-Variable Modeling (Hoyle et al. 2019)
- Reference: Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts (Merullo et al. 2019)
- Reference: Language (Technology) is Power: A Critical Survey of “Bias" in NLP (Lin Blodgett et al. 2020)
- Reference: Toward Gender-Inclusive Coreference Resolution (Trista Cao and Daume 2020)
- Reference: Reducing Gender Bias in Neural Machine Translation as a Domain Adaptation Problem (Saunders and Byrne 2020)
- Reference: Towards Controllable Biases in Language Generation (Sheng et al. 2020)
- Reference: Gender as a Variable in Natural-Language Processing: Ethical Considerations (Larson 2017)
- Reference: Do Artifacts Have Politics? (Winner 1980)
- Reference: The Trouble With Bias (Crawford 2017)
- Reference: Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview (Shah et al. 2020)
- Reference: Moving beyond “algorithmic bias is a data problem" (Hooker 2021)
- Reference: Fairness and Machine Learning (Barocas et al. 2019)
Slides: Bias Slides
Video: Bias Video