Keyphrase Generation for Scientific Document Retrieval

TL;DR

This paper demonstrates that sequence-to-sequence keyphrase generation models can enhance scientific document retrieval and introduces an evaluation framework to analyze their limitations across different domains.

Contribution

It provides empirical evidence of the benefits of keyphrase generation for retrieval and proposes a new framework for evaluating these models' limitations.

Findings

Keyphrase models improve retrieval performance.

Challenges exist in generating absent keyphrases.

Cross-domain generalization remains difficult.

Abstract

Sequence-to-sequence models have lead to significant progress in keyphrase generation, but it remains unknown whether they are reliable enough to be beneficial for document retrieval. This study provides empirical evidence that such models can significantly improve retrieval performance, and introduces a new extrinsic evaluation framework that allows for a better understanding of the limitations of keyphrase generation models. Using this framework, we point out and discuss the difficulties encountered with supplementing documents with -- not present in text -- keyphrases, and generalizing models across domains. Our code is available at https://github.com/boudinfl/ir-using-kg