On Mitigating Gender Bias in Natural Language Models


Authors : Aryan Gupta

Volume/Issue : Volume 6 - 2021, Issue 9 - September

Google Scholar : http://bitly.ws/gu88

Scribd : https://bit.ly/3uNmZ5n

As the world accelerates towards digitization, natural language generation (NLG) is becom ing a critical ingredient of common AI systems like Amazon’s Alexa and Apple’s Siri. How- ever, many recent studies have highlighted that machine learning models employed in NLG often inherit and amplify the societal biases in data – including gender bias. This paper aims to achieve gender parity in natural lan- guage models by analyzing and mitigating gen- der bias. An open-source corpus has been used to train and fine-tune the GPT-2 model, following which text is generated from prompts to investigate and mitigate the bias. Domain Adaptive Pretraining is used as the primary technique to counter the bias and the paper evaluates its effectiveness in contrast to other methods. Lastly, the impact of domain adaptation on the performance of the natural lan- guage model is looked at through perplexity of the de-biased model obtained. Through em- pirical and in-depth assessment of gender bias, this study provides a foundation for ameliorat- ing gender equality in the digitalspace

CALL FOR PAPERS


Paper Submission Last Date
31 - December - 2022

Paper Review Notification
In 1-2 Days

Paper Publishing
In 2-3 Days

Video Explanation for Published paper

Never miss an update from Papermashup

Get notified about the latest tutorials and downloads.

Subscribe by Email

Get alerts directly into your inbox after each post and stay updated.
Subscribe
OR

Subscribe by RSS

Add our RSS to your feedreader to get regular updates from us.
Subscribe