The project is leveraging AI technologies such as machine learning and deep learning to develop
a system for predicting deforestation. The system development is primarily being carried out by members of the National Institute of Advanced Industrial Science and Technology (AIST), but the goal is for the Brazilian side to maintain and improve the system sustainably in the future. To achieve this, it is essential to acquire the skills necessary not only for predictive modeling but also for applying AI in other fields. Therefore, the second training in Japan, focused on acquiring AI skills, was conducted from August 26 to September 6.
This year’s training welcomed the same five counterparts who participated last year. Building on the knowledge gained through regular study sessions with long-term experts since the previous training, this year’s program emphasized hands-on programming. During the training, prototypes were developed for programs that detect mining areas from satellite data and monitor deforestation from mining land changes where illegal logging has occurred. Based on these prototypes, the counterparts plan to develop fully-fledged AI programs upon returning to Brazil.
Additionally, there were active discussions with the AIST team on methods to improve the accuracy of predictive models in the future.