# Projective Methods for Mitigating Gender Bias in Pre-trained Language Models

arXiv: https://arxiv.org/abs/2403.18803 · HTML: https://mlsanity.com/paper/2403.18803/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 41 references that are in the map.
- First posted on arXiv 27 Mar 2024.

## Builds on (most cited first, 23 of 23)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [A large annotated corpus for learning natural language inference](https://mlsanity.com/paper/1508.05326/) (2015)
- [Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings](https://mlsanity.com/paper/1607.06520/) (2016)
- [Language (Technology) is Power: A Critical Survey of "Bias" in NLP](https://mlsanity.com/paper/2005.14050/) (2020)
- [StereoSet: Measuring stereotypical bias in pretrained language models](https://mlsanity.com/paper/2004.09456/) (2020)
- [A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity](https://mlsanity.com/paper/2302.04023/) (2023)
- [CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models](https://mlsanity.com/paper/2010.00133/) (2020)
- [On Measuring Social Biases in Sentence Encoders](https://mlsanity.com/paper/1903.10561/) (2019)
- [Mitigating Gender Bias in Natural Language Processing: Literature Review](https://mlsanity.com/paper/1906.08976/) (2019)
- [Measuring Bias in Contextualized Word Representations](https://mlsanity.com/paper/1906.07337/) (2019)
- [Learning Gender-Neutral Word Embeddings](https://mlsanity.com/paper/1809.01496/) (2018)
- [Intrinsic Bias Metrics Do Not Correlate with Application Bias](https://mlsanity.com/paper/2012.15859/) (2020)
- [On Measuring and Mitigating Biased Inferences of Word Embeddings](https://mlsanity.com/paper/1908.09369/) (2019)
- [Recent Advances in Natural Language Inference: A Survey of Benchmarks, Resources, and Approaches](https://mlsanity.com/paper/1904.01172/) (2019)
- [Are ChatGPT and GPT-4 General-Purpose Solvers for Financial Text Analytics? A Study on Several Typical Tasks](https://mlsanity.com/paper/2305.05862/) (2023)
- [Can NLI Models Verify QA Systems' Predictions?](https://mlsanity.com/paper/2104.08731/) (2021)
- [Evaluating Semantic Accuracy of Data-to-Text Generation with Natural Language Inference](https://mlsanity.com/paper/2011.10819/) (2020)
- [Investigating Gender Bias in BERT](https://mlsanity.com/paper/2009.05021/) (2020)
- [A General Framework for Implicit and Explicit Debiasing of Distributional Word Vector Spaces](https://mlsanity.com/paper/1909.06092/) (2019)
- [Marked Attribute Bias in Natural Language Inference](https://mlsanity.com/paper/2109.14039/) (2021)
- [Sequential Neural Networks for Noetic End-to-End Response Selection](https://mlsanity.com/paper/2003.02126/) (2020)

## Abstract (arXiv)

Mitigation of gender bias in NLP has a long history tied to debiasing static word embeddings. More recently, attention has shifted to debiasing pre-trained language models. We study to what extent the simplest projective debiasing methods, developed for word embeddings, can help when applied to BERT's internal representations. Projective methods are fast to implement, use a small number of saved parameters, and make no updates to the existing model parameters. We evaluate the efficacy of the methods in reducing both intrinsic bias, as measured by BERT's next sentence prediction task, and in mitigating observed bias in a downstream setting when fine-tuned. To this end, we also provide a critical analysis of a popular gender-bias assessment test for quantifying intrinsic bias, resulting in an enhanced test set and new bias measures. We find that projective methods can be effective at both intrinsic bias and downstream bias mitigation, but that the two outcomes are not necessarily correlated. This finding serves as a warning that intrinsic bias test sets, based either on language modeling tasks or next sentence prediction, should not be the only benchmark in developing a debiased language model.