My work focuses on understanding convolutional neural network (CNN) feature spaces. Most recently, we have learned that different networks learn (almost) the same feature space when trained on the same data. This research will be used in the DARPA Active Interpretation of Disparate Alternatives program. In addition, I contribute to the lab's work under the DARPA Communicating with Computers program.
- D. McNeely-White, R. Beveridge and B. Draper. Inception and Resnet: Same Training, Same Features, Proceedings of Behaviorally Inspired Cognitive Architectures 2019, Seattle, WA, Aug. 16-18, 2019. Springer-Verlag, A. Samsonovich (ed).
- S. Mukherjee, N. Cain, J. Walker, D. White, I. Ray and I. Ray. POSTER: PenJ1939: An Interactive Framework for Design and Dissemination of Exploits for Commercial Vehicles. 24th ACM SIGSAC Conference on Computer and Communications Security (CCS) 2017, Dallas, TX, Oct. 30-Nov 3, 2017.
- O. Rios, W. Tenhaeff, M. McGuire, P. Menchhofer, A. Johs, K. More, D. White Carbon Nanotube Enhanced Functional Carbon Fibers from Renewable Resources. Honolulu PRiME 2012, Oct. 7-12, 2012.