Have you ever had a language model, like Google translate, translate the Swedish word ‘sin’ as ‘his’ even though it should be ‘her’, or only gotten stories about men from ChatGPT?
Although many language technologies exhibit behaviors we consider overtly sexist, these are not the only ways that they participate in human systems of oppression such as sexism, racism, and transphobia.
This talk will cover how language technologies contribute to our understanding of “gender”, as well as some of the social implications of such contributions. We will also explore some of the possible approaches to understanding the “biases” in language technologies from a feminist and queer standpoint.
Hannah is affiliated with both the Department of Computing Science and Umeå Centre for Gender Studies, where they recently successfully defended their PhD thesis, titled "Gender and Representation: Investigations of Bias in Natural Language Processing" (svensk översättning: Genus och representation : studier av social bias i språkteknologi.) They hold a bachelor's degree in Computer Science, Linguistics, and French from Trinity College Dublin, and a Master of Science degree in Speech and Language Processing from the University of Edinburgh.
Their research focuses on understanding the power structures and biases at play in the types of language data used to train language models, as well as the outputs produced by such models.
If you would like to have breakfast, please register through this link. Registration deadline 21 May, 10.00
https://forms.office.com/e/TZjnPVYwZe