Deep learning techniques are making major advances in solving problems that have resisted the best attempts of the artificial intelligence (A.I.) community to date, including record breaking image recognition across millions of images and thousands of classes. These advances are driving research, development, and commercial
implementations from major technology companies and suppliers. The integrated MBARI time series project (901512) has identified a need for rapid classification of key species represented in large time series data sets to drive model development for carbon flow in Monterey Bay. Guided by goals of the ITS science team, we propose to develop a limited, well defined prototype by mid 2016 to evaluate these systems to perform rapid classification of identified key species, capitalizing on research achievements in A.I. as well as affordable, turn-key hardware and software implementations.
Through such activities we hope to identify high risk areas, promising new opportunities, and to provide technical guidance to these various work teams as we navigate through a new era of exploration and discovery.