AI-assisted Algal Occurrence & Trait Database

The project aims to develop an AI-assisted database integrating algal occurrence and functional trait information from scientific publications, reports and existing datasets. A locally hosted Large Language Model (LLM) will be used to extract and standardize taxonomic, spatial, temporal and ecological information, together with morphological and functional traits. The database will also incorporate advanced trait descriptors, including projected area, cell and colony dimensions, geometric characteristics and 3D morphology. All extracted information will be linked to its original source and subjected to automated taxonomic and quality-control procedures. GPU resources are required for local LLM inference and batch processing of large document collections. The resulting PostgreSQL/PostGIS-based system will support ecological research, biodiversity assessment, phytoplankton functional analyses and Water Framework Directive-related applications.