Education
PhD at Tel Aviv University’s School of Zoology
Experience
Shir Bar is a postdoctoral fellow in the Beery Lab at MIT CSAIL. Her research focuses on adapting recent advances in artificial intelligence and machine learning to the messy, data-limited, and highly contextual challenges of ecological research. Trained as a marine ecologist, Shir has a longstanding love for the underwater world and for fishes as fascinating, diverse, and often challenging study systems. During her MSc and PhD, she conducted research at the Interuniversity Institute for Marine Sciences in Eilat, studying fish ecology in the wild, on the reef, and in aquaculture mesocosms. After encountering major analysis bottlenecks, she turned toward computer vision as a way to study fish behavior in situ, focusing on behaviors that are hard to observe, rare, or difficult to define in advance. Shir’s current research extends some of these ideas into movement ecology, where she is developing lightweight, human-in-the-loop tools to help ecologists work through large GPS tracking datasets and review questionable movement segments in context. More broadly, she is interested in how practical AI tools can help researchers find small but important events in large ecological datasets, from data errors to rare behaviors.
Having spent countless hours manually analyzing reef imagery — and leaving many videos unanalyzed — Shir is passionate about teaching computer vision to ecologists and helping researchers scale their analyses in ways that remain grounded in biology.
Selected publications
Roberts, C., Vergés, A., Bar, S., Sharma, T. and Poore, A., 2026. Trawling for photographic bycatch: using computer vision to detect habitat-forming kelp in citizen science fish photographs. MEPS 2026. [paper]
Bar S, Hirschorn O, Holzman R, Avidan S. Sifting through the haystack-efficiently finding rare animal behaviors in large-scale datasets. WACV 2025. [paper]
Bar, S., Levy, L., Avidan, S. and Holzman, R. Assessing the determinants of larval fish strike rates using computer vision. Ecological Informatics 2023 [paper]
