Synthesizing Human Gaze Feedback for Improved NLP Performance

TL;DR

This paper introduces ScanTextGAN, a model that synthesizes human gaze patterns to enhance NLP models, reducing the need for costly eye-tracking data and improving task performance.

Contribution

We propose ScanTextGAN, a novel generative model for creating human-like scanpaths over text, enabling the use of synthetic gaze data in NLP tasks.

Findings

Generated scanpaths approximate real human gaze patterns.

Augmenting NLP models with synthetic scanpaths improves performance.

Synthetic gaze data can substitute expensive eye-tracking data.

Abstract

Integrating human feedback in models can improve the performance of natural language processing (NLP) models. Feedback can be either explicit (e.g. ranking used in training language models) or implicit (e.g. using human cognitive signals in the form of eyetracking). Prior eye tracking and NLP research reveal that cognitive processes, such as human scanpaths, gleaned from human gaze patterns aid in the understanding and performance of NLP models. However, the collection of real eyetracking data for NLP tasks is challenging due to the requirement of expensive and precise equipment coupled with privacy invasion issues. To address this challenge, we propose ScanTextGAN, a novel model for generating human scanpaths over text. We show that ScanTextGAN-generated scanpaths can approximate meaningful cognitive signals in human gaze patterns. We include synthetically generated scanpaths in four popular NLP tasks spanning six different datasets as proof of concept and show that the models augmented with generated scanpaths improve the performance of all downstream NLP tasks.