Entity Tracking Improves Cloze-style Reading Comprehension

TL;DR

This paper enhances reading comprehension models by incorporating entity tracking through additional features and multi-task training, significantly improving performance on challenging entity-related questions.

Contribution

It introduces simple yet effective extensions for entity tracking in neural models, outperforming previous state-of-the-art methods on the LAMBADA dataset.

Findings

Improved accuracy on entity-related questions

Enhanced model performance with entity features and multi-task training

Outperforms previous state-of-the-art on LAMBADA dataset

Abstract

Reading comprehension tasks test the ability of models to process long-term context and remember salient information. Recent work has shown that relatively simple neural methods such as the Attention Sum-Reader can perform well on these tasks; however, these systems still significantly trail human performance. Analysis suggests that many of the remaining hard instances are related to the inability to track entity-references throughout documents. This work focuses on these hard entity tracking cases with two extensions: (1) additional entity features, and (2) training with a multi-task tracking objective. We show that these simple modifications improve performance both independently and in combination, and we outperform the previous state of the art on the LAMBADA dataset, particularly on difficult entity examples.