Masterarbeit
AI-Assisted Knowledge Graph Construction for Deep Learning Models
Research Area
Intelligent Information Management
Advisers
Dr. Sheeba Samuel
Description
The rapid growth of Artificial Intelligence and Deep Learning has led to the publication of thousands of machine learning models across scientific papers and open repositories. Structured metadata and semantic technologies, such as Knowledge Graphs, provide a promising approach to organize and connect information about research artifacts. At the same time, modern AI techniques, including large language models and automated information extraction methods, enable the extraction of structured information from unstructured sources such as repository documentation or research papers. Combining these approaches can support the creation of machine-readable knowledge about deep learning models. The objective of this thesis is to design a structured metadata schema for describing deep learning models and to develop an AI-assisted approach for constructing a knowledge graph from information available in open data and software repositories. An AI-based extraction pipeline will be implemented to analyze information from open repositories and map the extracted data to the proposed schema. Finally, the structured metadata will be used to build a knowledge graph that links models, datasets, repositories, and publications.
