As a follow-up to the Memorandum of Understanding (MoU) between Yayasan Ekosistem Lestari (YEL) and Universitas Syiah Kuala, signed on March 2, 2026, by the two leaders, Yayasan Ekosistem Lestari and the Data Science and Artificial Intelligence Research Center (DSAI-RC) Universitas Syiah Kuala agree to sign an Implementation Arrangement to enhance the GeoFlux model and platform by utilizing AI technology and multi-modal data sources.
The activity to enhance the GeoFlux model and platform from a disaster-monitoring repository into a multimodal data-source flood prediction system will include three main phases, starting with Data Acquisition and Processing. In this implementation phase, both teams will collect and process flood information from multiple data sources using AI technology, including documents, news articles, webpages, regulations, and other related sources. The information will be transformed into a structured, machine-readable format while maintaining traceability to its sources. The model will analyze the data, map affected areas, and identify recurring flood events. The second phase is the Micro-Scale Deforestation Analysis. In this phase, the teams will leverage advanced satellite embedding models to detect small-scale forest cover loss across the critical Leuser and Ulu Masen watersheds that drives downstream flood vulnerability. Last but not least is Automatic Speech-to-Text Recognition, which uses Acehnese and Gayo speech-to-text language models to convert disaster audio into Acehnese and Gayo text.
The implementation will run from September 15, 2026, to October 30, 2027.
