Eye-tracking data enhances keyphrase extraction from academic texts

Researchers have developed a novel approach to keyphrase extraction by incorporating human reading behavior, specifically eye-tracking data. They created a cost-effective webcam-based system and a Chinese academic eye-tracking corpus (CLIS-ET) for library and information science abstracts. The study found that features like first fixation duration, fixation number, and total fixation duration consistently improved keyphrase extraction performance, demonstrating the value of integrating reader attention signals into text analysis. AI

IMPACT This research could lead to more nuanced and accurate keyphrase extraction systems by incorporating human cognitive signals.

RANK_REASON The item is an academic paper detailing a new methodology and dataset for keyphrase extraction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• alphaXiv
• Att-BiLSTM+CRF
• CatalyzeX Code Finder for Papers
• CLIS-ET
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Influence Flower
• library and information science
• ScienceCast
• SearchGazer

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →