A Cross-Domain Transferable Neural Coherence Model

TL;DR

This paper introduces a neural coherence model that effectively generalizes across different text domains by using a local discriminative approach with reduced negative sampling, outperforming previous methods.

Contribution

A novel local discriminative neural coherence model that enhances cross-domain transferability and outperforms existing models on standard and new challenging datasets.

Findings

Significantly outperforms previous state-of-the-art methods.

Effective in transfer to unseen categories of discourse.

Simpler structure with efficient learning against incorrect orderings.

Abstract

Coherence is an important aspect of text quality and is crucial for ensuring its readability. One important limitation of existing coherence models is that training on one domain does not easily generalize to unseen categories of text. Previous work advocates for generative models for cross-domain generalization, because for discriminative models, the space of incoherent sentence orderings to discriminate against during training is prohibitively large. In this work, we propose a local discriminative neural model with a much smaller negative sampling space that can efficiently learn against incorrect orderings. The proposed coherence model is simple in structure, yet it significantly outperforms previous state-of-art methods on a standard benchmark dataset on the Wall Street Journal corpus, as well as in multiple new challenging settings of transfer to unseen categories of discourse on Wikipedia articles.