NEW YORK, February 4, 2011 -- A defining feature of a scientific discovery is replication by others. In today's age of computational science, this means higher standards of communication of discoveries -- making available the data that generated the results along with the published research paper. Doing this makes the technology behind the finding widely accessible, facilitating re-use and verification of results.
Tools and approaches to facilitate such knowledge transfer will be discussed at a symposium titled The Digitization of Science: Reproducibility and Interdisciplinary Knowledge Transfer at the American Association for the Advancement of Science Annual Meeting in Washington, DC, on Saturday, February 19, 1:30 p.m. to 4:30 p.m., in 159AB Washington Convention Center.
Victoria Stodden, Columbia University, symposium organizer
Keith Baggerly, M.D. Anderson Cancer Research Center
David Donoho, Stanford University
Matan Gavish, Stanford University
Robert Gentleman, University of Washington
Mark Liberman, University of Pennsylvania
Michael Reich, Broad Institute of Harvard and MIT
Fernando Perez, University of California at Berkeley
Included among the range of new tools for presentation at the symposium are Donoho and Gavish's Universal Identifier for Computational Results which creates a registry for computational results that provides replication and citation information; Stodden's Reproducible Research Standard, a suite of open licenses to bring the intellectual property framework faced by computational scientists in line with longstanding scientific norms; Baggerly's case study of the widely discussed 2010 terminated clinical trials at Duke occurring in part because of efforts by his lab to reproduce the associated studies; and Liberman's technical communication methodologies learned from the DARPA Speech and Language Program.
The AAAS symposium highlights how such knowledge transfer can detect flawed science and the importance of reproducibility for error control. Open code and data permits computational science to achieve standards for reproducibility that characterize the scientific endeavor, and maximize the wide availability of scientific knowledge.