CyberInfrastructure for video analysis of individual bee behavior
The goal of the CyIndiBee project is to build innovative cyberinfrastructure showcasing the use of modern AI approaches for transforming the study of pollinators behavior down to the individual level. It leverages the latest advances in machine-learning to enable both fine-grained and long-term analyses of behavior, enabling innovative studies in climate change research.
It relies on 4 thrusts:
Thrust 1) AI models leveraging Vision Transformers to overcome biological data scarcity and complexity
Thrust 2) Integrated platform bridging edge computing and HPC/cloud computing to collect in the field and process behavioral datasets at scale.
Thrust 3) Data Analysis Tools to reconstruct individual foraging activity enabling the discovery of long-term patterns
Thrust 4) Interdisciplinary co-design loop between computer science and biology experts
The models and tools developed enable real-time behavior monitoring in the field as well as offline analysis for long-term behavior.
