Oral Candidacy - Haiyan Meng
|Start:||10/5/2015 at 12:00PM|
|End:||10/5/2015 at 2:30PM|
|Location:||257G Fitzpatrick Hall|
Faculty and students are welcome to attend the presentation portion of the Oral Candidacy Exam.
October 5, 2015 257G Fitzpatrick Hall 12:00 pm
Dr. Douglas Thain
Dr. Sandra Gesing Dr. Michael Hildreth Dr. Jaroslaw Nabrzyski
Two Broad Approaches For Facilitating Reproducible Computational Science
Reproducibility is a cornerstone of scientific work, which has been guaranteed by text and proofs in a publication historically. However, as computation pervades the science and changes the scientific method, only the experimental results in a publication can not guarantee the reproducibility anymore. The execution environment, in which the results are generated, must also be preserved to reproduce the results. Unfortunately, execution environments for scientific work are often fragile, and too complex to be well understood by the researchers, let alone to be preserved.
To facilitate the preservation of scientific work and improve the reproducibility of scientific results, I propose two broad approaches for facilitating reproducible computational science - preserve the mess and encourage cleanliness, and explore their feasibility and applicability for single-machine applications and complex scientific workflows. For each approach, I plan to explore the following three aspects: what to preserve, how to preserve, and how to reproduce. A prototype will be implemented for each approach, and the time and space overhead to preserve, the time and space overhead to reproduce, the correctness of preserved artifacts, and the easiness to use of each prototype will be evaluated through applications from high energy physics, bioinformatics and biometrics.