Uncertainty Quantification Working Group
August 23, 11:30 AM, CNLS Conf. Room

Eliciting and representing expertsí conceptual knowledge

Laura McNamara and Jerry Morzinski (D-1)

Abstract

In order to make sense of data, all researchers must have a conceptual framework that allows them to contextualize and interpret "raw" information. Careful elicitation and mapping of these interpretive frameworks can help communities of researchers identify, quantify, and mitigate sources of uncertainty.

This talk introduces methods from cultural anthropology for eliciting and modeling expertsí knowledge, including sources of uncertainty. Using examples from our own experience with multidisciplinary research teams, we present methods for formally eliciting expert judgment from subject matter experts, for documenting expert judgment and developing representations for the subject matter expertsí review and refinement, and for combining expert judgment with other information sources. Not infrequently, differences of opinion arise among researchers, and an important aspect of elicitation is learning how to cope with such differences.

To send e-mail to authors: lauramc@lanl.gov or morzinski@lanl.gov