Connecting learned remote-sensing- and inSitu-based features to dynamic inverse reinforcement learning for interpretable cognition and movement modeling at scales ranging from individual animals to a given ecosystem. Aiming to provide a basis for modeling in marine and terrestrial environmnents.
Currently linking with https://github.com/nlahaye/SIT_FUSE for Remote-Sensing-based representation learning, but is built to be somewhat source agnostic.
Plans to add connections to https://github.com/SimonDedman/MarSpatAuto for marine insitu-based feature incorporation as a subsequent step.