stefano Fiorini

Dr. stefano Fiorini

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 June 2023 in Proceedings of the AAAI Conference on Artificial Intelligence
Verified
Journal Article
Open access
PDF file
SFstefano Fiorini
stefano Fiorini
SCStefano Coniglio
Stefano Coniglio
MCMichele Ciavotta
Michele Ciavotta
EMEnza Messina
Enza Messina
This paper introduces SigMaNet, a generalized Graph Convolutional Network (GCN) capable of handling both undirected and directed graphs with weights not restricted in sign nor magnitude. The cornerstone of SigMaNet is the Sign-Magnetic Laplacian (LSM), a new Laplacian matrix that we introduce ex novo in this work. LSM allows us to bridge a gap in the current literature by extending the theory of spectral GCNs to (directed) graphs with both positive and negative weights. LSM exhibits several desirable properties not enjoyed by other Laplacian matrices on which several state-of-the-art architectures are based, among which encoding the edge direction and weight in a clear and natural way that is not negatively affected by the weight magnitude. LSM is also completely parameter-free, which is not the case of other Laplacian operators such as, e.g., the Magnetic Laplacian. The versatility and the performance of our proposed approach is amply demonstrated via computational experiments. Indeed, our results show that, for at least a metric, SigMaNet achieves the best performance in 15 out of 21 cases and either the first- or second-best performance in 21 cases out of 21, even when compared to architectures that are either more complex or that, due to being designed for a narrower class of graphs, should---but do not---achieve a better performance.

14 October 2022 in Algorithms
Verified
Journal Article
Open access
PDF file
SFstefano Fiorini
stefano Fiorini
MCMichele Ciavotta
Michele Ciavotta
AMAndrea Maurino
Andrea Maurino
In recent years, studying and predicting mobility patterns in urban environments has become increasingly important as accurate and timely information on current and future vehicle flows can successfully increase the quality and availability of transportation services (e.g., sharing services). However, predicting the number of incoming and outgoing vehicles for different city areas is challenging due to the nonlinear spatial and temporal dependencies typical of urban mobility patterns. In this work, we propose STREED-Net, a novel autoencoder architecture featuring time-distributed convolutions, cascade hierarchical units and two distinct attention mechanisms (one spatial and one temporal) that effectively captures and exploits complex spatial and temporal patterns in mobility data for short-term flow prediction problem. The results of a thorough experimental analysis using real-life data are reported, indicating that the proposed model improves the state-of-the-art for this task.

30 July 2022 in EPJ Data Science
Verified
Journal Article
Open access
PDF file
SFstefano Fiorini
stefano Fiorini
MCMichele Ciavotta
Michele Ciavotta
SJSagar Joglekar
Sagar Joglekar
SŠSanja Šćepanović
Sanja Šćepanović
DQDaniele Quercia
Daniele Quercia
Recent years have witnessed the emerging of novel shared mobility solutions that provide diffused on-demand access to transportation. The widespread adoption of these solutions, particularly electric mopeds (e-mopeds), is expected to bring important benefits such as the reduction of noise and atmospheric pollution, and road congestion, with extensive repercussions on liveability and quality of life in urban areas. Currently, almost no effort has been devoted to exploring the adoption patterns of e-moped sharing services, therefore, optimal management and allocation of vehicles appears to be a problem for service managers. In this study, we tried to demonstrate the validity of the hypothesis that the adoption of electric mopeds depends on the built environment and demographic aspects of each neighbourhood. In detail, we singled out three features concerning the area characteristics (distance from centre, walkability, concentration of places) and one about the population (education index). The results obtained on a real world case study show the strong impact these factors have in determining the adoption of e-moped sharing services. Finally, an analysis was conducted on the possible role that the electric moped sharing can play in social equalization by studying the interactions between rich and poor neighbourhoods. The results of the analyses conducted indicate that communities within a city tend to aggregate by wealth and isolate themselves from one another (social isolation): very few interactions, in terms of trajectories, have been observed between the richest and poorest areas of the city under study.

18 March 2021
Verified
Book Chapter
Closed access
PDF file
EKElson Kurian
Elson Kurian
SVSherwin Varghese
Sherwin Varghese
SFstefano Fiorini
stefano Fiorini

11 September 2026
Verified
Book Chapter
Closed access
PDF file
AIAdeela Islam
Adeela Islam
SFstefano Fiorini
stefano Fiorini
MLManuel Lecha
Manuel Lecha
TTTheodore Tsesmelis
Theodore Tsesmelis
SJStuart James
Stuart James
ABAlessio Del Bue
Alessio Del Bue

2 January 2023 in arXiv
Verified
Preprint
Open access
PDF file
SFstefano Fiorini
stefano Fiorini
SCStefano Coniglio
Stefano Coniglio
MCMichele Ciavotta
Michele Ciavotta
EMEnza Messina
Enza Messina
This paper introduces SigMaNet, a generalized Graph Convolutional Network (GCN) capable of handling both undirected and directed graphs with weights not restricted in sign nor magnitude. The cornerstone of SigMaNet is the Sign-Magnetic Laplacian ($L^{\sigma}$), a new Laplacian matrix that we introduce ex novo in this work. $L^{\sigma}$ allows us to bridge a gap in the current literature by extending the theory of spectral GCNs to (directed) graphs with both positive and negative weights. $L^{\sigma}$ exhibits several desirable properties not enjoyed by other Laplacian matrices on which several state-of-the-art architectures are based, among which encoding the edge direction and weight in a clear and natural way that is not negatively affected by the weight magnitude. $L^{\sigma}$ is also completely parameter-free, which is not the case of other Laplacian operators such as, e.g., the Magnetic Laplacian. The versatility and the performance of our proposed approach is amply demonstrated via computational experiments. Indeed, our results show that, for at least a metric, SigMaNet achieves the best performance in 15 out of 21 cases and either the first- or second-best performance in 21 cases out of 21, even when compared to architectures that are either more complex or that, due to being designed for a narrower class of graphs, should -- but do not -- achieve a better performance.