ICArUS: Phase 1 Test and Evaluation Development Guide

The Integrated Cognitive-neuroscience Architectures for Understanding Sensemaking (ICArUS) Program aimed to build computational cognitive-neuroscience models to explain, predict, and emulate the process of human sensemaking – the process of generating hypotheses to explain data that is sparse, noisy, and uncertain. Phase 1 lasted from December, 2010 to December, 2012. In Phase 1, the program employed a challenge problem, based on geospatial data, to compare the models’ performance to human performance. Here, the term geospatial data includes multiple types of intelligence data, any of which may be represented in a georeferenced framework – for instance, a layer within a GIS-like display. The challenge problem was focused on scientific rigor as opposed to operational relevance; all examples were fictitious in nature and represented in a neutral manner. This document’s primary purpose is to specify the format of the ICArUS Phase 1 Challenge Problem. It was written for software developers, contains development code, and is not meant to stand alone.