I develop and apply machine learning techniques, informed by (mostly marine) ecological questions and engineering constraints, to facilitate observing systems that address long-standing scientific questions with societal relevance. I lead the AI team at the UK National Oceanography Centre where we work on diverse problems, from decision making on autonomus systems to analyzing global-scale model output.

News

• 2025-07-24 Our paper on using out-of-the-box foundation models to assist annotators of marine imagery is out in Frontiers.
• 2025-06-30 Welcome to Polina, our summer intern through the IGNITE DLA program! Polina will collect plankton samples from NOC’s pontoon, develop a computer vision model to classify the images, and examine population changes.
• 2025-05-24 Jonathan gave an introduction to our work in the ML for Ocean Modelling workshop at the University of Reading.
• 2025-06-27 Alex and Jonathan’s work with the NERC funded API4AI project resulted in a white paper on making UK enviromental data AI-ready.
• 2025-05-07 Applications are now open for the 2026 Computer Vision for Ecology Short Course. This year we are be hosting the three week class at the Smithsonian Conservation Biology Insitute in the lovely Blue Ridge Mountains.
• 2025-03-28 Our paper Machine Learning Techniques to Characterize Functional Traits of Plankton from Image Data was one of Limnology and Oceanography’s most read articles in 2024. We were recongized by an invitation to present at ASLO’s Spotlight Session at Aquatic Sciences.
• 2025-01-04 I am co-leading the Workshop on Computer Vision Methods for Ecology workshop at the Caltech Resnick Sustainability Institute with Sara Beery.
• 2024-06-18 We presented early results from the FathomNet2024 competition at the 11th Workshop on Fine-Grained Visual Categorization at CVPR.
• 2024-04-05 Our paper on tracking-by-detection methods to enable target acquisition and following on underwater vehicles is out in Frontiers.
• 2024-01-17 Our paper on digital holotyping resulting from Schmidt Ocean’s Designing the Future Two cruise was published in Science Advances.
• 2024-01-15 PAIDIVER, our project developing software to support estimating biodiversity metrics from image data, has been funded by the NERC AIAB program!
• 2023-12-16 Dani Ocampo presented our workshop paper on VLM failure modes on images of organisms at the I Can’t Believe It’s Not Better Workshop at NeurIPS 2023.
• 2023-12-15 I was an invited panelist at the Computational Sustainability workshop at NeurIPS 2023.