IST LUNCH BUNCH

"Enabling Reproducibility in Computational Science"

 

Imagine querying the scholarly record for all image compression algorithms that have been applied to the famous "Barbara" image in the last five years (with citations), or all articles published using the well-known Acute Lymphoblastic Leukemia dataset from Golub et al. (1999). Queries such as these are natural questions for researchers, and are as yet effectively impossible. The scholarly community is taking steps to link data, code, workflows and other artifacts that support and enable the verification of the claims made in the scholarly record. In this talk I will frame a motivation for this effort -- resolving reproducibility in computational science -- and discuss recent steps to add this layer to the scholarly record including a recently funded NSF project, "Merging Science and Cyberinfrastructure Pathways: The Whole Tale".