Enhancing Unsupervised Keyword Extraction in Academic Papers through Integrating Highlights with Abstract

2026-04-21 · Source: Computation and Language · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Intermediate, quick

Summary

This paper explores enhancing unsupervised keyword extraction from academic papers by integrating the highlights section with the abstract. While prior research primarily used abstracts and references, this study focuses on highlights, which summarize key findings and contributions. Observations indicate highlights offer valuable keyword information that complements abstracts. The research evaluates three input scenarios: abstract only, highlights only, and a combination of both, using four unsupervised models on Computer Science (CS) and Library and Information Science (LIS) datasets. Experiments demonstrate that combining abstracts with highlights significantly improves keyword extraction performance. The study also examines how differences in keyword coverage and content between abstracts and highlights influence extraction outcomes. Data and code are publicly available.

Key takeaway

For research scientists developing information retrieval systems for academic literature, you should consider integrating the "highlights" section of papers with their abstracts. This approach has been shown to significantly improve unsupervised keyword extraction performance, leading to more accurate and comprehensive indexing. Incorporating this combined input could enhance the discoverability and utility of research papers within your systems.

Key insights

Integrating academic paper highlights with abstracts significantly improves unsupervised keyword extraction performance.

Principles

• Highlights complement abstracts for keyword extraction.
• Combined inputs outperform single inputs for keyword extraction.

Method

Evaluated three input scenarios (abstract, highlights, combined) with four unsupervised models on CS and LIS datasets to measure keyword extraction performance.

In practice

• Combine abstract and highlights for better keyword extraction.
• Apply unsupervised models to academic paper analysis.

Topics

• Keyword Extraction
• Academic Papers
• Highlights Section
• Abstract Integration
• Unsupervised Models

Code references

• xiangyi-njust/Highlight-KPE

Best for: Research Scientist, AI Scientist, Machine Learning Engineer, NLP Engineer

Related on AIssential

• Website Content Analysis: Keyword and Terminology Extraction. Implementation on Apify — Automated keyword and terminology extraction from websites aids content analysis, SEO, and trend identification.
• Which Sections of a Research Paper Best Reveal Its Research Methods? Evidence from Library and Information Science — The most effective research method classification comes from middle-to-late and final paper segments combined with bibliographic metadata.
• Attention Expansion: Enhancing Keyphrase Extraction from Long Documents with Attention-Augmented Contextualized Embeddings — Attention expansion efficiently broadens PLM context for keyphrase extraction by integrating lightweight out-of-context word embeddings.
• ScholarSum: Student-Teacher Abstractive Summarization via Knowledge Graph Reasoning and Reflective Refinement — ScholarSum uses a student-teacher, graph-based approach to achieve fluent and faithful scientific abstractive summarization.
• A Comparative Study of Multi-document Summarization Techniques — The paper compares multi-document summarization techniques, contributing to NLP research.
• Beyond Keyword Alerts: Building a Semantic PR Opportunity Monitor for a Real Agency — Semantic retrieval and RAG effectively identify proactive PR opportunities beyond traditional keyword matching.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computation and Language.