Topic-Aware Neural Keyphrase Generation for Social Media Language

TL;DR

This paper introduces a topic-aware seq2seq neural model for social media keyphrase generation, capable of producing absent keyphrases and improving performance by modeling latent topics, demonstrated on multilingual datasets.

Contribution

It presents a novel topic-aware neural keyphrase generation framework that addresses data sparsity and enhances keyphrase prediction for social media content.

Findings

Outperforms existing extraction and generation models on multiple datasets

Learns meaningful latent topics that improve keyphrase generation

Effectively generates absent keyphrases in social media language

Abstract

A huge volume of user-generated content is daily produced on social media. To facilitate automatic language understanding, we study keyphrase prediction, distilling salient information from massive posts. While most existing methods extract words from source posts to form keyphrases, we propose a sequence-to-sequence (seq2seq) based neural keyphrase generation framework, enabling absent keyphrases to be created. Moreover, our model, being topic-aware, allows joint modeling of corpus-level latent topic representations, which helps alleviate the data sparsity that widely exhibited in social media language. Experiments on three datasets collected from English and Chinese social media platforms show that our model significantly outperforms both extraction and generation models that do not exploit latent topics. Further discussions show that our model learns meaningful topics, which interprets its superiority in social media keyphrase generation.