AI & Climate Advocacy

A Committee of FXB Climate Advocates

Project: Biodiversity Preservation through AI

Biodiversity is very important to the health of the earth’s ecosystem, and we want to preserve it as best as possible. We want to save the animal species that are affected by the changing climate before it’s too late. It is critical to measure and monitor biodiversity in a systematic and ongoing basis as part of wildlife management and conservation. Camera traps that record wildlife with heat and motion sensors play important role in such monitoring. Despite its limitations, camera trap data has been a very important source of information used for tracking biodiversity.


However, such data, millions of images, takes a lot of data curation. Manual processing times cause significant delays between data collection and ecological action. Machine learning models can reduce the processing times, but are not as frequently used because users lack confidence in the model’s precision and accuracy.


We aim to develop creative ways to improve Convolutional Neural Network (CNN) training, that enables us to quickly learn from the past misclassifications and errors to enhance such models’ performance, precision, and accuracy. We also seek to facilitate standardization of metadata format globally to enhance data sharing and explore innovative ways of ongoing and systematic biodiversity monitoring.