Project data

The primary goal of this project is the advancement of the research of deep neural networks, in the field of adaptive processing of graph data (Deep Graph Learning, DGL). Graph structures naturally characterize a wide range of problems, including the areas of bio/chemistry, natural language processing, social network, object identification. By being able to deal with the inherent nature of structured data, learning models are endowed with a formidable capability and flexibility both to address new domains and to improve accuracy and efficiency in solving complex problems. The project includes the following strongly interconnected fundamental research topics: (i) introduction of highly efficient DGL models to reduce the energy impact and increase the sustainability of DGL models; (ii) increase the expressiveness of DGL models, obtaining better predictive performances on existing tasks and enabling the application of DGL in novel tasks where current methods do not achieve satisfying performances; (iii) extending the scope of DGL application, not only in terms of the considered tasks but also on the considered setting: for instance, dynamic (spatio/temporal) graphs pose several challenges and are not well studied to date, as well.

 

The project is funded by Italian Ministry of University and Research under the PRIN program – Progetti di Rilevante Interesse Nazionale (PRIN) – PRIN 2022 (Secretary-General’s Decree n. 1401 del 18/09/2024) – CUP B53C24006570006