Logo image
Back
Empirical Analysis of Multi-Task Learning for Reducing Model Bias in Toxic Comment Detection
Preprint   Open access

Empirical Analysis of Multi-Task Learning for Reducing Model Bias in Toxic Comment Detection

Ameya Vaidya, Feng Mai and Yue Ning
ArXiv.org
Cornell University
03/27/2020
DOI: 10.48550/arxiv.1909.09758
url
https://doi.org/10.48550/arxiv.1909.09758View
Preprint (Author's original)This preprint has not been evaluated by subject experts through peer review. Preprints may undergo extensive changes and/or become peer-reviewed journal articles. Open Access

Abstract

With the recent rise of toxicity in online conversations on social media platforms, using modern machine learning algorithms for toxic comment detection has become a central focus of many online applications. Researchers and companies have developed a variety of models to identify toxicity in online conversations, reviews, or comments with mixed successes. However, many existing approaches have learned to incorrectly associate non-toxic comments that have certain trigger-words (e.g. gay, lesbian, black, muslim) as a potential source of toxicity. In this paper, we evaluate several state-of-the-art models with the specific focus of reducing model bias towards these commonly-attacked identity groups. We propose a multi-task learning model with an attention layer that jointly learns to predict the toxicity of a comment as well as the identities present in the comments in order to reduce this bias. We then compare our model to an array of shallow and deep-learning models using metrics designed especially to test for unintended model bias within these identity groups.
Computer Science - Artificial Intelligence

Details

Metrics

22 Record Views
Logo image