Show, Write, and Retrieve: Entity-aware Article Generation and Retrieval

TL;DR

The paper introduces ENGINE, a framework that explicitly incorporates named entities into language models to improve article generation and retrieval, especially for news stories referencing real-world entities.

Contribution

ENGINE is the first framework to explicitly integrate named entities into language models for improved article generation and retrieval tasks.

Findings

Boosted article generation perplexity by 4-5 points.

Improved article retrieval recall@1 by 3-4%.

Effective on three public datasets: GoodNews, VisualNews, WikiText.

Abstract

Article comprehension is an important challenge in natural language processing with many applications such as article generation or image-to-article retrieval. Prior work typically encodes all tokens in articles uniformly using pretrained language models. However, in many applications, such as understanding news stories, these articles are based on real-world events and may reference many named entities that are difficult to accurately recognize and predict by language models. To address this challenge, we propose an ENtity-aware article GeneratIoN and rEtrieval (ENGINE) framework, to explicitly incorporate named entities into language models. ENGINE has two main components: a named-entity extraction module to extract named entities from both metadata and embedded images associated with articles, and an entity-aware mechanism that enhances the model's ability to recognize and predict entity names. We conducted experiments on three public datasets: GoodNews, VisualNews, and WikiText, where our results demonstrate that our model can boost both article generation and article retrieval performance, with a 4-5 perplexity improvement in article generation and a 3-4% boost in recall@1 in article retrieval. We release our implementation at https://github.com/Zhongping-Zhang/ENGINE .