Masterarbeit / Bachelorarbeit
Continuous Reproducibility Service for Scientific Publications
Research Area
Advisers
Dr. Sheeba Samuel
Description
In modern scientific research, ensuring the reproducibility of computational results is a growing challenge, particularly as publications increasingly rely on code repositories and interactive notebooks. Many articles in PubMed Central reference resources hosted on platforms such as GitHub or Jupyter Notebook, yet the information needed to reproduce experiments is often scattered or incomplete. As a result, researchers struggle to identify, retrieve, and execute the required code and environments for reproducing published results. This thesis aims to design and implement a web-based reproducibility service that continuously collects publications, extracts reproducibility-related metadata, and links them to their associated repositories and computational artifacts. The system will automatically detect repository references, clone the code, and prepare execution environments for reproducible experiments. An intuitive web portal will allow users to search publications, inspect extracted metadata, and initiate reproduction workflows. The outcome of the thesis will be a prototype platform demonstrating how AI-assisted metadata extraction and repository mining can support continuous reproducibility in scientific research.
