Collaborative: EAGER: Exploring and Advancing the State of the Art in Robust Science in Gravitational Wave Physics

University Of Tennessee Knoxville, Knoxville TN

Investigators

Abstract

Science is increasingly based on computation for science simulations, data management and analysis, instrument control and collaboration. For scientific results generated through computation to be considered robust and become widely accepted, the computational techniques should be automated, reproducible and trustworthy. By exploring the practices of gravitational-wave astronomy researchers working on the Laser Interferometer Gravitational-Wave Observatory (LIGO) project, this project seeks to create a set of case studies documenting broadly applicable methods for reproducible computational science. Specifically, the project will explore and articulate what reproducibility, automation, and trust mean with respect to computation-based research in gravitational-wave astronomy, identify, implement and validate a set of experimental practices, that will include computational techniques, and finally, evaluate how these experimental practices can be extended to other science domains. Robust computational science builds on rigorous methods and is composed of three key elements: (1) reproducibility, which enables the verification and leveraging of scientists' findings; (2) automation, which speeds up the exploration of alternative solutions and the processing of large amounts of data while reducing the introduction of errors; and (3) trust, providing security and reliability for software and data, while supplying the necessary attributes for confidence in the scientist's own results and results from others. This project explores robust science in the LIGO project through the following activities within the context of gravitational-wave astronomy: (1) articulating the roles of reproducibility, automation, and trust in gravitational-wave astronomy; (2) identifying, implementing and validating a set of experimental practices, including computational techniques; and (3) advancing towards the project's vision of general computational methods for robust science by evaluating how the experimental practices can be extended to other science domains. The project will develop and use a survey to collect information about LIGO workflows that are composed of a series of experimental, computational, and data manipulation steps. The analysis of the survey will result in a document that describes what reproducibility means in the LIGO context and help identify potential improvements in LIGO's practices. The project will generalize these findings by documenting a mapping of LIGOÕs original and enhanced approach to other science workflows including those of the molecular dynamics and bioinformatics communities. The final project document will target a broad audience that includes researchers and students at various levels of education, with the goal of introducing them to the concept of robust computational research, and the underlying concepts of reproducibility, automation and trust, teaching them to access code, data, and workflow information to regenerate findings, learn about the scientific methods, and to engage in STEM 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.