The Grossman–Cormack Glossary of Technology-Assisted Review
Federal Courts Law Review, Vol. 7 (2013)
A common vocabulary for a field in which the same term was being used for different things—and different terms for the same thing.
Overview
The glossary was created to bring order to the rapidly developing vocabulary of technology-assisted review. It defines the central concepts of machine learning, information retrieval, statistics, sampling, and legal document review in language intended for the bench, the bar, service providers, and technical practitioners.
Key contributions
- Distinguishes TAR from narrower labels such as predictive coding.
- Defines recall, precision, richness, sampling, training, control sets, and related concepts.
- Connects legal review vocabulary with established information-retrieval and statistical terminology.
- Includes a foreword by U.S. Magistrate Judge John M. Facciola.
- Was explicitly cited by courts as an explanation of TAR technology and terminology.
Paper summary
The introduction of TAR brought new vernacular and considerable confusion. The glossary addresses cases in which different terms referred to the same process, the same terms referred to different processes, and technical concepts were misunderstood or distorted. Its purpose is to provide a consistent framework and concise definitions for judges, lawyers, service providers, and researchers.
Citation
Maura R. Grossman & Gordon V. Cormack, The Grossman–Cormack Glossary of Technology-Assisted Review, with Foreword by John M. Facciola, U.S. Magistrate Judge, 7 Fed. Cts. L. Rev. 1 (2013).