ICI Bucharest participated in the International Summer School “The Use of Digital Twin and Artificial Intelligence Tools in Aquaculture”, held in Constanța from 7 to 11 September 2026. As part of the programme, the institute organised, through the BSC AI Factory project, the “Experiment Tracking for Computer Vision” training session, a practical course designed to help participants better organise their experiments, compare different model variants, and obtain results that are easier to track and reproduce.
The session focused on the use of MLflow and Weights & Biases, two tools commonly used to monitor Machine Learning experiments. Participants learned how to keep key experiment information in one place, from hyperparameters and metrics to models, images, and predictions.
The training included demonstrations and hands-on exercises using MLflow and Weights & Biases, covering everything from logging and comparing multiple runs to analysing images, predictions, confusion matrices, and learning curves. The summer school programme combined theoretical presentations with practical and participatory activities. These included satellite data analysis, sampling and water quality measurements at Siutghiol Lake, biological and bacterial characterisation activities, as well as data collection in the Black Sea. The Black Sea Universities Network (BSUN) organised the summer school within the Blue Gates – Empowering Blue and Smart Transformation of the Black Sea Basin project.
The event brought together researchers, academics, and experts interested in using emerging technologies to develop more efficient and sustainable solutions for aquatic ecosystems and aquaculture.