Development of Domain-Specific Scenarios for Training and Evaluation of Two-Way, Free Form, Spoken Language Translation Devices
Brian A. Weiss, Marnie Menzel · 2010
To create effective and accurate two-way, free form, spoken language translation devices, the technologies must have appropriate training data. The goal of the DARPA (Defense Advanced Research Projects Agency) TRANSTAC (Spoken Language Communication and Translation System for Tactical Use) program is to demonstrate capabilities to rapidly develop and field this technology so speakers of different languages can communicate in real-world tactical situations. A critical component is to generate data sets to train and evaluate the technologies. A novel ap-proach was developed to collect these data employing innovative data collection and evaluation scenarios. This paper describes the scenario methodology used for the TRANSTAC data collec-tions and evaluations.