Powering climate change effects analysis with knowledge graphs and embeddings, a strategy for Latin America
Climate change is one of the greatest threats to the world today. Climate change affects countries without environmental culture, such as Bolivia, which suffer from recurrent problems such as forest fires, lack of access to clean water, high heat spots, and untimely droughts in various regions. In Bolivia, there are currently scattered sources of information regarding climate change. Although historical data on temperature and metadata related to the same scheme have been detected, there is no solution for taking preventive actions on climate precipitation and/or natural disasters, since such data work in isolation and without correlation.
Artificial intelligence (AI) can enhance deeper and more accurate research on disaster prediction and climate change in Bolivia, but an analysis of its applicability in the Bolivian context must be carried out. In this project, we propose the review of available datasets at the national level for integration using Knowledge Graphs, unifying heterogeneous data into actionable knowledge for analysis, labeling, and use in Graph Neural Networks (GNN). Missing data relevant to predictive models that Bolivia needs to generate will be reviewed. Finally, a long-term action plan will be developed using KG and GNN, with which we will describe key performance indicators and thus proceed with the review of possible scenarios in which our technology will be useful and of high environmental impact, not only in Bolivia but in any country in the region.
TEAM
Diego Collarana, Ph.D. (Main Contact)
E-mail: diegocollarana@upb.edu
Carla Marcela Florida Román
E-mail: carlaflorida@upb.edu
Fernanda Lorena Martínez Moreno
E-mail: fernandamartinez1@upb.edu
Erika Fernandez Lujan
