CICI: TCR: Improving Computational Research Reproducibility

Scientific reproducibility remains difficult in practice, particularly for research in computer science that depends on complex software environments, evolving dependencies, specialized hardware, and undocumented execution steps. This project develops a framework that captures, reconstructs, and publicly re-executes scientific experiments to make computational results easier to reproduce and verify. To do this, the team is integrating software supply chain technologies with scientific workflows to automatically record software artifacts, execution environments, system dependencies, and experimental actions during normal research use. Researchers invoke the framework through a simple command-line interface while otherwise conducting experiments as usual. The resulting metadata and artifacts are being indexed in a large-scale repository. Experiments are reconstructed in cloud-based execution environments that support independent verification and long-term reproducibility. This work will lower the barriers to computational reproducibility across scientific domains and help make reproducible experimentation a routine part of scientific practice. The resulting system will support more reliable scientific results, provide broader community validation, and leverage cloud providers to improve long-term preservation of computational research. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.