Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods
arXiv:1804.06876
Abstract
We introduce a new benchmark, WinoBias, for coreference resolution focused on gender bias. Our corpus contains Winograd-schema style sentences with entities corresponding to people referred by their occupation (e.g. the nurse, the doctor, the carpenter). We demonstrate that a rule-based, a feature-rich, and a neural coreference system all link gendered pronouns to pro-stereotypical entities with higher accuracy than anti-stereotypical entities, by an average difference of 21.1 in F1 score. Finally, we demonstrate a data-augmentation approach that, in combination with existing word-embedding debiasing techniques, removes the bias demonstrated by these systems in WinoBias without significantly affecting their performance on existing coreference benchmark datasets. Our dataset and code are available at http://winobias.org.
arXiv commentNAACL '18 Camera Ready
References in corpus (2)
Cited by in corpus (32)
- Multitask Prompted Training Enables Zero-Shot Task Generalization
- WILDS: A Benchmark of in-the-Wild Distribution Shifts
- Learning the Difference that Makes a Difference with Counterfactually-Augmented Data
- Representation Bias in Data: A Survey on Identification and Resolution Techniques
- WinoGrande: An Adversarial Winograd Schema Challenge at Scale
- Gender Bias in Neural Natural Language Processing
- Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias
- Learning from others' mistakes: Avoiding dataset biases without modeling them
- Balanced Datasets Are Not Enough: Estimating and Mitigating Gender Bias in Deep Image Representations
- Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems
- Evaluating the Underlying Gender Bias in Contextualized Word Embeddings
- Man is to Person as Woman is to Location: Measuring Gender Bias in Named Entity Recognition
- Utility is in the Eye of the User: A Critique of NLP Leaderboards
- Intersectional Bias in Causal Language Models
- Undesirable Biases in NLP: Addressing Challenges of Measurement
- Online Abuse of UK MPs from 2015 to 2019: Working Paper
- Exposing and Correcting the Gender Bias in Image Captioning Datasets and Models
- Mind the GAP: A Balanced Corpus of Gendered Ambiguous Pronouns
- Explaining The Efficacy of Counterfactually Augmented Data
- Evaluating Gender Bias in Natural Language Inference
- A Causal Inference Method for Reducing Gender Bias in Word Embedding Relations
- LangFair: A Python Package for Assessing Bias and Fairness in Large Language Model Use Cases
- Race and Religion in Online Abuse towards UK Politicians: Working Paper
- Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionately
- Fair Hate Speech Detection through Evaluation of Social Group Counterfactuals
- Towards classification parity across cohorts
- MABR: Multilayer Adversarial Bias Removal Without Prior Bias Knowledge
- [RE] Double-Hard Debias: Tailoring Word Embeddings for Gender Bias Mitigation
- Perceptual Score: What Data Modalities Does Your Model Perceive?
- Enhancing Model Robustness and Fairness with Causality: A Regularization Approach
- Improving Counterfactual Generation for Fair Hate Speech Detection
- Small Business Classification By Name: Addressing Gender and Geographic Origin Biases