at 11 a.m.
In hybrid
Room COP-2165 (ULAVAL)
Room Estelle Laberge P-212 (ISMER)
 |
Eric ORENSTEIN, (Monterey Bay Aquarium Research Institute) |
Intelligent autonomous sampling of marine ecosystems
Ocean dynamics constantly influence the presence, abundance, and density of marine organisms. The resulting rapidly fluctuating, spatially heterogeneous biological distributions make measurement and monitoring an ongoing challenge. Finding and persistently sampling zones of elevated biological activity in the ocean is particularly vexing; these features are ephemeral and highly localized, challenging most conventional observation systems. Concurrent advances in three areas – in situ imaging systems, automated image analysis, and autonomous sampling platforms – promise to help scientists better understand how these ubiquitous structures shape community ecology. There are many avenues for autonomous targeted sampling to illuminate cryptic facets of the community, expand monitoring coverage, and provide new understanding of physical-biological interactions. But actually combining these tools to intelligently sample these complex ecosystems remains a difficult, and exciting, area of development. In this talk, I will give some historical perspective on marine imaging, on-going efforts to leverage AI to inform robotic behavior, and the complexities there-in.
Bio:
Eric Orenstein is a Research Engineer at the Monterey Bay Aquarium Research Institute (MBARI). His work lives at the intersection of machine learning, ocean imaging, and marine ecology. Automated detection and classification models often fail in the context of rapidly fluctuating ocean ecosystems and conditions. Eric spends a lot of time thinking about the root causes of that behavior and ways of ameliorating them to speed ecological analysis. Eric received his PhD at the Scripps Institution of Oceanography developing automated classification approaches to process in situ plankton image data. After completing his degree, he stayed on at Scripps and continued to leverage techniques from his thesis work to investigate cryptic host-parasite interactions and harmful algal bloom formation. Before arriving at MBARI, Eric worked at the Sorbonne Université in the Laboratoire d’Océanographie de Villefranche on automated techniques for extracting functional trait information from images of ocean organisms
