Deep Attentive Ranking Networks for Learning to Order Sentences

TL;DR

This paper introduces an attention-based ranking framework utilizing bidirectional encoders and self-attention transformers to improve sentence ordering and discrimination tasks, outperforming existing methods.

Contribution

It proposes a novel, input order invariant framework that supports multiple ranking loss functions, enhancing sentence ordering performance.

Findings

Outperforms state-of-the-art methods on sentence ordering and order discrimination tasks.

Pairwise and listwise ranking losses yield better results than pointwise loss.

The framework effectively captures relative sentence positions for improved learning.

Abstract

We present an attention-based ranking framework for learning to order sentences given a paragraph. Our framework is built on a bidirectional sentence encoder and a self-attention based transformer network to obtain an input order invariant representation of paragraphs. Moreover, it allows seamless training using a variety of ranking based loss functions, such as pointwise, pairwise, and listwise ranking. We apply our framework on two tasks: Sentence Ordering and Order Discrimination. Our framework outperforms various state-of-the-art methods on these tasks on a variety of evaluation metrics. We also show that it achieves better results when using pairwise and listwise ranking losses, rather than the pointwise ranking loss, which suggests that incorporating relative positions of two or more sentences in the loss function contributes to better learning.