End-Task Oriented Textual Entailment via Deep Explorations of Inter-Sentence Interactions

TL;DR

This paper introduces DEISTE, a novel deep learning model that enhances textual entailment detection by exploring inter-sentence interactions, significantly improving performance on the SciTail dataset and generalizing well to other entailment tasks.

Contribution

DEISTE employs dynamic convolution and position-aware attentive convolution to better model inter-sentence interactions for entailment detection, advancing the state of the art.

Findings

DEISTE achieves approximately 5% improvement over previous methods.

Pretrained DEISTE generalizes well to RTE-5 dataset.

Model effectively captures word-to-word interactions for entailment.

Abstract

This work deals with SciTail, a natural entailment challenge derived from a multi-choice question answering problem. The premises and hypotheses in SciTail were generated with no awareness of each other, and did not specifically aim at the entailment task. This makes it more challenging than other entailment data sets and more directly useful to the end-task -- question answering. We propose DEISTE (deep explorations of inter-sentence interactions for textual entailment) for this entailment task. Given word-to-word interactions between the premise-hypothesis pair ($P$, $H$), DEISTE consists of: (i) a parameter-dynamic convolution to make important words in $P$ and $H$ play a dominant role in learnt representations; and (ii) a position-aware attentive convolution to encode the representation and position information of the aligned word pairs. Experiments show that DEISTE gets $\approx$5\% improvement over prior state of the art and that the pretrained DEISTE on SciTail generalizes well on RTE-5.