Ben Dutton Email & Phone Number

Continuing work on disentangling information in multi-view learning environments.

Raleigh-Durham, North Carolina Area

I do research on ways to use remote sensing to detect nuclear facilities focusing on the invariance/discrimination tradeoff and designing algorithms for which invariance (usually, but not exclusively, to certain mathematical groups) can be user-specified.

Seattle

I design machine learning algorithms for early detection of nuclear activity for the purposes of non-proliferation.

Raleigh, North Carolina

I am currently working in Dr. Raju Vatsavai's STAC lab on an unmixing problem that seeks to identify endmembers that have certain symmetry properties.Last year, I worked with Dr. Chris Parnin on two problems: 1) learning word embeddings in the Stack Overflow dataset and 2) using these work embeddings to predict author gender.

Oak Ridge, Tn

Worked with the cyber security team to create a deep learning-based anomaly detection algorithm.

Oak Ridge, Tn

Did research on feature learning/selection for network intrusion detection problems.

Houston, Texas Area

Designed and implemented (R,S,s) inventory model that sets reorder points and drives purchasing decisions for a two-tier distribution system with stochastic lead times and demand.Designed reporting system used to govern the above inventory model that focuses on integrating information across large domains to make buying decisions faster and easier. The reporting system focuses on items where the model is most likely to fail. The previous system, lacking any framework that would allow such distinctions, required an exhaustive review of items. This system has allowed the purchasing to be done by 3 people instead of 7.Designed and implemented a separate (S,s) inventory model for high dollar, high volume, long lead-time items that relies on forecast error and lead-time variability to drive the safety stock. The volume allows for effective forecasting which means that, despite the long lead-times (2+ months), we can keep low safety stock, resulting in higher turns than for general low-volume items with shorter lead times.

Doctor Of Philosophy (Phd), Machine Learning, 4.0 Overall

Activities and Societies: Dean's Merit AwardCurrently working with Dr. Chris Parnin on applying deep learning techniques to the stack.

Bachelor Of Science (B.S.), Mathematics Minor Computer Science, 3.97 Overall

Activities and Societies: Pi Mu Epsilon, Golden Key Honor Society, Dean's List every semesterGraduated with highest honors (summa cum.

Bachelor Of Science (B.S.), Industrial Distribution, 3.75 Over Last 90 Credit Hours