Metadata Automation and Enrichment Using Machine Learning

This chapter examines how machine learning is transforming metadata automation and enrichment across research infrastructures, libraries, and data repositories. Synthesizing natural language processing, computer vision, large language models, and knowledge graph approaches, it analyzes how automated subject indexing, entity extraction, image captioning, author disambiguation, schema crosswalks, and ontology learning address persistent metadata quality deficits at scale. A machine learning enrichment pipeline model and a taxonomy of ML approaches are introduced, alongside a comparative modality table and performance bubble chart. The chapter critically examines FAIR data alignment, ethical governance, and implementation barriers.

Hali tarjima qilinmagan