Refining Raw Sentence Representations for Textual Entailment Recognition via Attention

TL;DR

This paper introduces a neural model that refines raw sentence representations with attention for textual entailment, achieving competitive accuracy on the RepEval shared task by combining LSTM encoding, aggregation, and attention mechanisms.

Contribution

The novel approach refines raw sentence representations with an attention mechanism, improving entailment recognition accuracy over baseline models.

Findings

Achieved over 72% accuracy on the RepEval dataset.

Outperformed the LSTM baseline in both tracks.

Ensemble methods further improved accuracy.

Abstract

In this paper we present the model used by the team Rivercorners for the 2017 RepEval shared task. First, our model separately encodes a pair of sentences into variable-length representations by using a bidirectional LSTM. Later, it creates fixed-length raw representations by means of simple aggregation functions, which are then refined using an attention mechanism. Finally it combines the refined representations of both sentences into a single vector to be used for classification. With this model we obtained test accuracies of 72.057% and 72.055% in the matched and mismatched evaluation tracks respectively, outperforming the LSTM baseline, and obtaining performances similar to a model that relies on shared information between sentences (ESIM). When using an ensemble both accuracies increased to 72.247% and 72.827% respectively.