I am Antonio Longa, a Postdoctoral Researcher at UiT The Arctic University of Norway, where I work with the Northernmost GraphML Group under the supervision of Filippo Maria Bianchi.
I am also a member of ELLIS.
Previously, I was an Assistant Professor at the University of Trento, working in the Structured Machine Learning Group led by Andrea Passerini.
I earned my Ph.D. with honors from the University of Trento and Fondazione Bruno Kessler, working in the MobS Lab under the supervision of Bruno Lepri.
My research focuses on Graph Machine Learning, Temporal Graphs, Relational Deep Learning, and Complex Systems, with interests in both theoretical foundations and real-world applications, including energy systems and materials science.
I am always happy to discuss science and research, so feel free to get in touch :)
Short research notes, tutorials, and explainers on Graph Machine Learning, Temporal Graphs, Network Science, and related topics.
Link Representation · Expressive GNNs
Link representation with graph neural networks: from understanding theoretical expressiveness, to comparing subgraph-based and heuristic methods, to bridging practical link prediction with principled multi-node representation learning.
Meta-Paths · Heterogeneous GNNs
Why meta-paths matter in heterogeneous graph neural networks: from MP-GNN, which learns informative relation chains automatically, to MPS-GNN, which extends them with statistics over repeated relational patterns.
Graph Transformers · Structural Encoding
A clear look at Simple Path Structural Encoding: why random-walk edge encodings can miss cyclic structure, how simple path counts improve graph transformer representations, and what the results show on molecular, long-range, and graph benchmark tasks.
Graph Regression · Software Performance
A benchmark for predicting Java execution time from source-code graphs, combining syntax, control flow, and data flow in homogeneous and multi-relational graph representations.
Temporal Graphs · Network Generation
A concise research story on egocentric temporal neighborhoods: from mining temporal motifs, to generating surrogate temporal networks, to modeling communities and higher-order interactions.
Chem Interactions
Interactive platform for exploring ligand–metal–solvent interactions and discovering MOFs using FAIR-MOF data-driven recommendations
Pytorch Geometric Tutorials
Collections of video tutorials and material for the implementation of GDL in PyG
Advanced Pytorch Geometric Tutorials
Collections of video tutorials and material for the implementation of advanced techniques in PyG
Egocentric Temporal Motifs Miner (ETMM)
Blog post presenting a novel mining technique based on the notion of Egocentric Temporal Motifs.
AI Trento Journal club
The AI Trento Journal Club is an open science initiative aiming to discuss and speculate about various topics of Machine Learning,
Artificial Intelligence, and Computer Science.
LOG meetup Trento
Trento meetup of the Learning on Graph Conference, an annual research conference that covers areas broadly related to machine learning on graphs and geometry.
Scientific Programming - Programming - 2025-2026
Master of Data Science - Trento University
Scientific Programming - Programming - 2024-2025
Master of Data Science - Trento University
Scientific Programming - Programming - 2023-2024
Master of Data Science - Trento University
Scientific Programming - Programming - 2022-2023
Master of Data Science - Trento University
Machine Learning - Module I - 2021 - 2022
Master of Science in Artificial Intelligence Systems - Trento University
Part 1 (Sklearn)
Machine Learning 2021 - 2022
Master of Science in Computer Science - Trento University
Part 1 (Sklearn) - Part 1 (Tensorflow)
Informatica 2020 - 2021
Scienze e Tecnologie Biomolecolari - Università degli Studi di Trento