Gender Bias Detection

We are working on studying methods to detect gendered language automatically using unsupervised learning methods, such as variational auto-encoders. The findings of our first paper on this (Hoyle et al., 2019) have been reported by 75+ international news outlets, including Forbes.

Currently, we’re interested in expanding the above to a cross-lingual study, as well as researching the relationship between gender bias and attitudes towards entities on social media as part of a project funded by DFF.

Moreover, in a new Carlsberg-funded project starting in autumn 2023, we’ll be investigating fair and accountable Natural Language Processing methods, which can be used to understand what influences the employer images that organisations project in job ads.

Publications

How much meaning influences gender assignment across languages is an active area of research in modern linguistics and cognitive …

Despite mounting evidence that women in foreign policy often bear the brunt of online hostility, the extent of online gender bias …

Large language models have been shown to encode a variety of social biases, which carries the risk of downstream harms. While the …

While the prevalence of large pre-trained language models has led to significant improvements in the performance of NLP systems, recent …

NLP models are used in a variety of critical social computing tasks, such as detecting sexist, racist, or otherwise hateful content. …

Pre-trained language models have been known to perpetuate biases from the underlying datasets to downstream tasks. However, these …

Despite attempts to increase gender parity in politics, global efforts have struggled to ensure equal female representation. This is …

Counterfactually Augmented Data (CAD) aims to improve out-of-domain generalizability, an indicator of model robustness. The improvement …

While the prevalence of large pre-trained language models has led to significant improvements in the performance of NLP systems, recent …

As NLP models are increasingly deployed in socially situated settings such as online abusive content detection, ensuring these models …

Machine Learning (ML) seeks to identify and encode bodies of knowledge within provided datasets. However, data encodes subjective …

Studying to what degree the language we use is gender-specific has long been an area of interest in socio-linguistics. Studies have …