A Parallel-Hierarchical Model for Machine Comprehension on Sparse Data

TL;DR

This paper introduces a neural parallel-hierarchical model for machine comprehension that effectively handles limited data, outperforming previous methods on the MCTest benchmark by leveraging multiple trainable perspectives.

Contribution

The work presents a novel neural architecture that compares text passages, questions, and answers from various perspectives, improving performance on small datasets.

Findings

Sets a new state-of-the-art on MCTest

Outperforms previous neural approaches by over 15%

Demonstrates effectiveness of multi-perspective neural comparison

Abstract

Understanding unstructured text is a major goal within natural language processing. Comprehension tests pose questions based on short text passages to evaluate such understanding. In this work, we investigate machine comprehension on the challenging {\it MCTest} benchmark. Partly because of its limited size, prior work on {\it MCTest} has focused mainly on engineering better features. We tackle the dataset with a neural approach, harnessing simple neural networks arranged in a parallel hierarchy. The parallel hierarchy enables our model to compare the passage, question, and answer from a variety of trainable perspectives, as opposed to using a manually designed, rigid feature set. Perspectives range from the word level to sentence fragments to sequences of sentences; the networks operate only on word-embedding representations of text. When trained with a methodology designed to help cope with limited training data, our Parallel-Hierarchical model sets a new state of the art for {\it MCTest}, outperforming previous feature-engineered approaches slightly and previous neural approaches by a significant margin (over 15\% absolute).