Albert Cabellos (PhD 2008) is a full professor since 2020 at the Computer Architecture Department (Universitat Politècnica de Catalunya). He is the co-founder of the Barcelona Neural Networking (https://bnn.upc.edu/) and the NaNoNetworking Center in Catalunya (https://www.n3cat.upc.edu/) . He has also founded the Open Overlay Router (http://openoverlayrouter.org) along with Cisco. He has been a visiting researcher at Cisco Systems and Agilent Technologies and a visiting professor at the Royal Institute of Technology (KTH), the Massachusetts Institute of Technology (MIT) and UC Berkeley. He has participated in several national (Cicyt), EU (H2020), USA (NSF) and industrial projects. He is the organizer of the Graph Neural Networking Challenge (https://bnn.upc.edu/challenge/). He has co-authored over 200 papers and participated in several IETF RFCs. 

Research interests

Graph Neural Network (GNN) is a new family of neural nets that have been proposed with the aim of learning graphs. GNNs facilitate the learning of relations between entities and show outstanding generalization capabilities. GNNs are not a black-box that simply map input to output labels from the dataset, as a consequence novel GNN architectures are required to tackle different problems. The introduction of Convolutional Neural Networks radically changed the field of computer vision. We argue that Graph Neural Networks will have a similar impact and represent a breakthrough for many fields that fundamentally operate with graphs. We plan to push progress on the theoretical foundations of Graph Neural Networks driven by relevant industrial use-cases.

Keywords

Computer Networks, Graph Neural Networks, Deep-Reinforcement Learning, Graphene, Nanotechnology