English
Related papers

Related papers: Unlabeled Debiasing in Downstream Tasks via Class-…

200 papers

This study explores using embedding rank as an unsupervised evaluation metric for general-purpose speech encoders trained via self-supervised learning (SSL). Traditionally, assessing the performance of these encoders is resource-intensive…

Audio and Speech Processing · Electrical Eng. & Systems 2025-01-22 Zakaria Aldeneh , Vimal Thilak , Takuya Higuchi , Barry-John Theobald , Tatiana Likhomanenko

Neural networks trained on biased datasets tend to inadvertently learn spurious correlations, hindering generalization. We formally prove that (1) samples that exhibit spurious correlations lie on a lower rank manifold relative to the ones…

Machine Learning · Computer Science 2024-11-07 Silpa Vadakkeeveetil Sreelatha , Adarsh Kappiyath , Abhra Chaudhuri , Anjan Dutta

We propose selective debiasing -- an inference-time safety mechanism designed to enhance the overall model quality in terms of prediction performance and fairness, especially in scenarios where retraining the model is impractical. The…

Computation and Language · Computer Science 2025-03-12 Gleb Kuzmin , Neemesh Yadav , Ivan Smirnov , Timothy Baldwin , Artem Shelmanov

When trained on large, unfiltered crawls from the internet, language models pick up and reproduce all kinds of undesirable biases that can be found in the data: they often generate racist, sexist, violent or otherwise toxic language. As…

Computation and Language · Computer Science 2021-09-10 Timo Schick , Sahana Udupa , Hinrich Schütze

Spurious correlations in training data significantly hinder the generalization capability of machine learning models when faced with distribution shifts, leading to the proposition of numberous debiasing methods. However, it remains to be…

Machine Learning · Computer Science 2025-05-22 Peng Kuang , Zhibo Wang , Zhixuan Chu , Jingyi Wang , Kui Ren

Although Large Language Models (LLMs) demonstrate remarkable reasoning capabilities, inherent social biases often cascade throughout the Chain-of-Thought (CoT) process, leading to continuous "Bias Propagation". Existing debiasing methods…

Computation and Language · Computer Science 2026-05-12 Xuan Feng , Shuai Zhao , Luwei Xiao , Tianlong Gu , Bo An

Test-time task adaptation in few-shot learning aims to adapt a pre-trained task-agnostic model for capturing taskspecific knowledge of the test task, rely only on few-labeled support samples. Previous approaches generally focus on…

Computer Vision and Pattern Recognition · Computer Science 2023-12-19 Ji Zhang , Lianli Gao , Xu Luo , Hengtao Shen , Jingkuan Song

The output tendencies of Pre-trained Language Models (PLM) vary markedly before and after Fine-Tuning (FT) due to the updates to the model parameters. These divergences in output tendencies result in a gap in the social biases of PLMs. For…

Computation and Language · Computer Science 2024-01-17 Masahiro Kaneko , Danushka Bollegala , Timothy Baldwin

Datasets often contain biases which unfairly disadvantage certain groups, and classifiers trained on such datasets can inherit these biases. In this paper, we provide a mathematical formulation of how this bias can arise. We do so by…

Machine Learning · Computer Science 2019-01-16 Heinrich Jiang , Ofir Nachum

Pre-trained language models have been successful on text classification tasks, but are prone to learning spurious correlations from biased datasets, and are thus vulnerable when making inferences in a new domain. Prior work reveals such…

Computation and Language · Computer Science 2022-01-03 Huihan Yao , Ying Chen , Qinyuan Ye , Xisen Jin , Xiang Ren

In the social sciences, it is often necessary to debias studies and surveys before valid conclusions can be drawn. Debiasing algorithms enable the computational removal of bias using sample weights. However, an issue arises when only a…

Machine Learning · Computer Science 2026-03-03 Tony Hauptmann , Stefan Kramer

Transfer learning is beneficial by allowing the expressive features of models pretrained on large-scale datasets to be finetuned for the target task of smaller, more domain-specific datasets. However, there is a concern that these…

Computer Vision and Pattern Recognition · Computer Science 2023-08-21 Angelina Wang , Olga Russakovsky

The task of image captioning implicitly involves gender identification. However, due to the gender bias in data, gender identification by an image captioning model suffers. Also, the gender-activity bias, owing to the word-by-word…

Computer Vision and Pattern Recognition · Computer Science 2019-12-03 Shruti Bhargava , David Forsyth

Recent generative large language models (LLMs) show remarkable performance in non-English languages, but when prompted in those languages they tend to express higher harmful social biases and toxicity levels. Prior work has shown that…

Computation and Language · Computer Science 2025-06-03 Vera Neplenbroek , Arianna Bisazza , Raquel Fernández

Large language models pick up social biases from the data they are trained on and carry those biases into downstream applications, often reinforcing stereotypes around gender, race, religion, disability, age, and socioeconomic status. The…

Computation and Language · Computer Science 2026-05-05 Muneeb Ur Raheem Khan

Recently, encoder-decoder neural models have achieved great success on text generation tasks. However, one problem of this kind of models is that their performances are usually limited by the scale of well-labeled data, which are very…

Computation and Language · Computer Science 2019-06-04 Hongyu Zang , Xiaojun Wan

Multi-modal Large Language Models (MLLMs) integrate visual and linguistic reasoning to address complex tasks such as image captioning and visual question answering. While MLLMs demonstrate remarkable versatility, MLLMs appears limited…

Computation and Language · Computer Science 2025-03-07 Wenke Huang , Jian Liang , Xianda Guo , Yiyang Fang , Guancheng Wan , Xuankun Rong , Chi Wen , Zekun Shi , Qingyun Li , Didi Zhu , Yanbiao Ma , Ke Liang , Bin Yang , He Li , Jiawei Shao , Mang Ye , Bo Du

When working with textual data, a natural application of disentangled representations is fair classification where the goal is to make predictions without being biased (or influenced) by sensitive attributes that may be present in the data…

Computation and Language · Computer Science 2022-10-10 Pierre Colombo , Guillaume Staerman , Nathan Noiry , Pablo Piantanida

Denoising diffusion models have emerged as a dominant approach for image generation, however they still suffer from slow convergence in training and color shift issues in sampling. In this paper, we identify that these obstacles can be…

Computer Vision and Pattern Recognition · Computer Science 2024-08-06 Hu Yu , Li Shen , Jie Huang , Hongsheng Li , Feng Zhao

Word embeddings are widely used in NLP for a vast range of tasks. It was shown that word embeddings derived from text corpora reflect gender biases in society. This phenomenon is pervasive and consistent across different word embedding…

Computation and Language · Computer Science 2019-09-25 Hila Gonen , Yoav Goldberg