Publications

When we are faced with challenging image classification tasks, we often explain our reasoning by dissecting the image, and pointing out prototypical aspects of one class or another. The mounting evidence for each of the classes helps us make our final decision. In this work, we introduce a deep network architecture that reasons in a similar way: the network dissects the image by finding prototypical parts, and combines evidence from the prototypes to make a final classification. The algorithm thus reasons in a way that is qualitatively similar to the way ornithologists, physicians, geologists, architects, and others would explain to people on how to solve challenging image classification tasks. The network uses only image-level labels for training, meaning that there are no labels for parts of images. We demonstrate the method on the CIFAR-10 dataset and 10 classes from the CUB-200-2011 dataset.

Artificial Intelligence (AI) has the opportunity to revolutionize the way the United States Department of Defense (DoD) and Intelligence Community (IC) address the challenges of evolving threats, data deluge, and rapid courses of action. Developing an end-to-end artificial intelligence system involves parallel development of different pieces that must work together in order to provide capabilities that can be used by decision makers, warfighters and analysts. These pieces include data collection, data conditioning, algorithms, computing, robust artificial intelligence, and human-machine teaming. While much of the popular press today surrounds advances in algorithms and computing, most modern AI systems leverage advances across numerous different fields. Further, while certain components may not be as visible to end-users as others, our experience has shown that each of these interrelated components play a major role in the success or failure of an AI system. This article is meant to highlight many of these technologies that are involved in an end-to-end AI system. The goal of this article is to provide readers with an overview of terminology, technical details and recent highlights from academia, industry and government. Where possible, we indicate relevant resources that can be used for further reading and understanding.

The Multi-modal Early Detection Interactive Classifier (MEDIC) is a triage system designed to enable rapid assessment of mild traumatic brain injury (mTBI) when access to expert diagnosis is limited as in a battlefield setting. MEDIC is based on supervised classification that requires three fundamental components to function correctly; these are data, features, and truth. The MEDIC system can act as a data collection device in addition to being an assessment tool. Therefore, it enables a solution to one of the fundamental challenges in understanding mTBI: the lack of useful data. The vision of MEDIC is to fuse results from stimulus tests in each of four modalitites - auditory, occular, vocal, and intracranial pressure - and provide them to a classifier. With appropriate data for training, the MEDIC classifier is expected to provide an immediate decision of whether the subject has a strong likelihood of having sustained an mTBI and therefore requires an expert diagnosis from a neurologist. The tests within each modalitity were designed to balance the capacity of objective assessment and the maturity of the underlying technology against the ability to distinguish injured from non-injured subjects according to published results. Selection of existing modalities and underlying features represents the best available, low cost, portable technology with a reasonable chance of success.