Title: A comprehensive survey on deep learning based keyphrase generation and extraction for natural language processing

Authors: Jimmy Jose; P. Beaulah Soundarabai

Addresses: CHRIST (Deemed to be University), Hosur Road, Near Dairy Circle, Bangalore, Karnataka-560029, India ' CHRIST (Deemed to be University), Hosur Road, Near Dairy Circle, Bangalore, Karnataka-560029, India

Abstract: Recent deep learning advances have revolutionised natural language processing through improved computational power and data availability. Keyphrases - condensed information representations - enhance performance in text summarisation, information retrieval, classification, sentiment analysis, and topic modelling. While previous surveys focused primarily on keyphrase extraction (KPE) and specific techniques, this comprehensive review uniquely analyses recent deep learning, embedding-based, and pre-trained language models in both keyphrase extraction and generation (KPG). The survey examines NLP advancements and their applications, with special focus on pre-trained language models like GPT and BERT and their impact on keyphrase tasks. It analyses strengths and limitations of recent KPG/KPE methods to help researchers improve existing approaches and addresses future research directions, ultimately promoting deeper understanding of KPG/KPE within the broader NLP landscape.

Keywords: natural language processing; NLP; keyphrase generation; KPG; keyphrase extraction; KPE; pre-trained language model; PLM; bidirectional encoder representations from transformers; BERT.

DOI: 10.1504/IJIIDS.2026.155303

International Journal of Intelligent Information and Database Systems, 2026 Vol.18 No.3/4, pp.411 - 455

Received: 01 Oct 2023
Accepted: 04 Nov 2024
Published online: 30 Jul 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article