MTLHealth: A Deep Learning System for Detecting Disturbing Content in Student Essays

TL;DR

This paper introduces MTLHealth, a deep learning system utilizing Transformer models to automatically detect disturbing content in student essays, aiding human graders in identifying potentially harmful references.

Contribution

The paper presents a novel deep learning pipeline specifically designed for detecting disturbing content in student essays, leveraging recent advances in pre-trained language models.

Findings

Effective detection of disturbing content in essays.

Improved accuracy over traditional keyword-based methods.

Supports human decision-making in educational settings.

Abstract

Essay submissions to standardized tests like the ACT occasionally include references to bullying, self-harm, violence, and other forms of disturbing content. Graders must take great care to identify cases like these and decide whether to alert authorities on behalf of students who may be in danger. There is a growing need for robust computer systems to support human decision-makers by automatically flagging potential instances of disturbing content. This paper describes MTLHealth, a disturbing content detection pipeline built around recent advances from computational linguistics, particularly pre-trained language model Transformer networks.