AI Test, Evaluation, and Assurance

Kitware has significant experience in the test and evaluation (T&E) of AI systems. We develop software to accelerate and enable AI model testing and evaluation across the DoD and provide insight into performance, effectiveness, robustness, and safety of the DoD’s AI-enabled systems. Kitware’s Explainable AI Toolkit (XAITK) can be used to explore, quantify, and monitor the behavior of deep learning systems. Kitware’s Natural Robustness Toolkit (NRTK) can be used to evaluate the robustness of deep neural networks when faced with previously-unknown conditions.These capabilities have been developed on a variety of research programs over the past ten years, funded by DARPA and CDAO.