A new paper from the University of Oxford explores the use of AI for keyword extraction from crowdsourced collections, using the Their Finest Hour Online Archive as a case study. The research evaluated three Natural Language Processing approaches: Named Entity Recognition, Keyword Extraction, and Topic Modelling, testing various AI techniques from statistical methods to generative AI. Findings suggest that while NLP methods show promise for large-scale keyword extraction, no single method is sufficient, and model choice impacts results. The paper also highlights the ethical stewardship responsibilities associated with automated keyword extraction from collections involving living contributors, favoring open-weight, extractive models for responsible deployment over generative AI due to accountability risks. AI
IMPACT Highlights the need for careful consideration of ethical implications and model choice when applying AI to crowdsourced data.
RANK_REASON The cluster contains an academic paper detailing research findings on AI techniques for keyword extraction.
- arXiv
- generative artificial intelligence
- keyword extraction
- Miguel Arana-Catania
- named-entity recognition
- natural language processing
- Their Finest Hour Online Archive
- topic modeling
- University of Oxford
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
