MDERank: A Masked Document Embedding Rank Approach for Unsupervised Keyphrase Extraction

TL;DR

MDERank introduces a novel unsupervised keyphrase extraction method that uses masked document embeddings and a specialized BERT model, significantly improving performance on multiple benchmarks.

Contribution

The paper proposes MDERank, a new unsupervised KPE approach utilizing masked embeddings and a contrastively trained KPEBERT model, addressing long document representation issues.

Findings

MDERank outperforms existing unsupervised methods by 1.80 F1@15 on average.

KPEBERT enhances embedding quality for keyphrase extraction.

Overall, MDERank achieves a 3.53 F1@15 improvement over SOTA methods.

Abstract

Keyphrase extraction (KPE) automatically extracts phrases in a document that provide a concise summary of the core content, which benefits downstream information retrieval and NLP tasks. Previous state-of-the-art (SOTA) methods select candidate keyphrases based on the similarity between learned representations of the candidates and the document. They suffer performance degradation on long documents due to discrepancy between sequence lengths which causes mismatch between representations of keyphrase candidates and the document. In this work, we propose a novel unsupervised embedding-based KPE approach, Masked Document Embedding Rank (MDERank), to address this problem by leveraging a mask strategy and ranking candidates by the similarity between embeddings of the source document and the masked document. We further develop a KPE-oriented BERT (KPEBERT) model by proposing a novel self-supervised contrastive learning method, which is more compatible to MDERank than vanilla BERT. Comprehensive evaluations on six KPE benchmarks demonstrate that the proposed MDERank outperforms state-of-the-art unsupervised KPE approach by average 1.80 $F1@15$ improvement. MDERank further benefits from KPEBERT and overall achieves average 3.53 $F1@15$ improvement over the SOTA SIFRank. Our code is available at \url{https://github.com/LinhanZ/mderank}.