paper-with-me

Papers

Debiasing Pretrained Text Encoders by Paying Attention to Paying Attention

2021-09-29 · Yacine GACI, Boualem Benatallah, Fabio Casati, Khalid Benabdeslem

Recent studies in fair Representation Learning have observed a strong inclination for natural language processing (NLP) models to exhibit discriminatory stereotypes across gender, religion, race and many such social constructs. In comparison to the progress made in reducing bias from static word embeddings, fairness in sentence-level text encoders received little consideration despite their wider applicability in contemporary NLP tasks. In this paper, we propose a debiasing method for pre-trained text encoders that both reduces social stereotypes, and inflicts next to no semantic offset. Unlike previous studies that directly manipulate the embeddings, we suggest to dive deeper into the operation of these encoders, and pay more attention to the way they pay attention to different social groups. We find that the attention mechanism is the root of all stereotypes. Then, we work on model debiasing by redistributing the attention scores of a text encoder such that it forgets any preference to historically advantaged groups, and attends to all social classes with the same intensity. Our experiments confirm that we successfully reduce bias with little damage to semantic representation.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

FairnessRepresentation LearningSentenceWord Embeddings

Similar Papers 제목 키워드 기반

FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders

2021-03-11 · ICLR 2021 1 · Pengyu Cheng, Weituo Hao, Siyang Yuan, Shijing Si 외

Pretrained text encoders, such as BERT, have been applied increasingly in various natural language processing (NLP) tasks, and have recently demonstrated significant performance gains. However, recent studies have demons…

Contrastive LearningFairnessSentence

FairNVT: Improving Fairness via Noise Injection in Vision Transformers

2026-04-18 · Qiaoyue Tang, Sepidehsadat Hosseini, Mengyao Zhai, Thibaut Durand 외 arxiv

This paper presents FairNVT, a lightweight debiasing framework for pretrained transformer-based encoders that improves both representation and prediction level fairness while preserving task accuracy. Unlike many existin…

FAIRT2V: Training-Free Debiasing for Text-to-Video Diffusion Models

2026-01-28 · Haonan Zhong, Wei Song, Tingxu Han, Maurice Pagnucco 외 arxiv

Text-to-video (T2V) diffusion models have achieved rapid progress, yet their demographic biases, particularly gender bias, remain largely unexplored. We present FairT2V, a training-free debiasing framework for text-to-vi…

Text-to-Video Generation

FineDeb: A Debiasing Framework for Language Models

2023-02-05 · Akash Saravanan, Dhruv Mullick, Habibur Rahman, Nidhi Hegde

As language models are increasingly included in human-facing machine learning tools, bias against demographic subgroups has gained attention. We propose FineDeb, a two-phase debiasing framework for language models that s…

Language ModelingLanguage Modelling

General Phrase Debiaser: Debiasing Masked Language Models at a Multi-Token Level

2023-11-23 · Bingkang Shi, Xiaodan Zhang, Dehan Kong, Yulei Wu 외

The social biases and unwelcome stereotypes revealed by pretrained language models are becoming obstacles to their application. Compared to numerous debiasing methods targeting word level, there has been relatively less …