Loïck Bonniot, industrial Phd student in the research team WIDE, and his team 'Steredeg' have won 1st place at the Graph Neural Networking Challenge 2020, organized as part of the AI/ML in 5G Challenge of l'ITU (the International Telecommunication Union), a specialized agency of the United Nations responsible for all matters related to information and communication technologies.
Loïck's solution uses a combination of Graph Neural Networks, recurrent neural networks, and aggregation strategies to predict the latency performance of complex network infrastructures with a 98.5% accuracy. Salzburg Research Forschungsgesellschaft mbH, and Fraunhofer HHI have come in second and third place, respectively.
Besides Loïck, the 'Steredeg' team also includes François Schnitzler and Christoph Neumann from InterDigital, and François Taïani Professor Université de Rennes 1 and member of the research team WIDE