2025
Journal article
Open Access
The path is the goal: a study on the nature and effects of shortest-path stability under perturbation of destination
Cornacchia G., Nanni M.
This work examines the phenomenon of path variability in urban navigation, where small changes in destination might lead to significantly different suggested routes. Starting from an observation of this variability over the city of Barcelona, we explore whether this is a localized or widespread occurrence and identify factors influencing path variability. We introduce the concept of “path stability”, a measure of how robust a suggested route is to minor destination adjustments, define a detailed experimentation process and apply it across multiple cities worldwide. Our analysis shows that path stability is shaped by city-specific factors and trip characteristics, also identifying some common patterns. Results reveal significant heterogeneity in path stability across cities, allowing for categorization into “path-stable” and “path-unstable” cities. These findings offer new insights for urban planning and traffic management, highlighting opportunities for optimizing navigation systems to enhance route consistency and urban mobility.Source: GEOINFORMATICA, vol. 29 (issue 4), pp. 975-998
DOI: 10.1007/s10707-025-00553-z
DOI: 10.48550/arxiv.2506.09731
Project(s): Green.Dat.AI
Metrics:
Cornacchia G., Nanni M.
This work examines the phenomenon of path variability in urban navigation, where small changes in destination might lead to significantly different suggested routes. Starting from an observation of this variability over the city of Barcelona, we explore whether this is a localized or widespread occurrence and identify factors influencing path variability. We introduce the concept of “path stability”, a measure of how robust a suggested route is to minor destination adjustments, define a detailed experimentation process and apply it across multiple cities worldwide. Our analysis shows that path stability is shaped by city-specific factors and trip characteristics, also identifying some common patterns. Results reveal significant heterogeneity in path stability across cities, allowing for categorization into “path-stable” and “path-unstable” cities. These findings offer new insights for urban planning and traffic management, highlighting opportunities for optimizing navigation systems to enhance route consistency and urban mobility.Source: GEOINFORMATICA, vol. 29 (issue 4), pp. 975-998
DOI: 10.1007/s10707-025-00553-z
DOI: 10.48550/arxiv.2506.09731
Project(s): Green.Dat.AI
Metrics:
See at:
arXiv.org e-Print Archive 
| GeoInformatica 
| doi.org | GitHub | CNR IRIS | CNR IRIS | CNR IRIS | link.springer.com
2025
Journal article
Open Access
Vehicle-pedestrian optimization framework for exposure-aware routing
Aliyev G., Nanni M.
Vehicular traffic is a major source of air pollution in urban areas, exposing pedestrians and residents to harmful emissions. Recent works have proposed exposure-aware pedestrian routing strategies based on static emission maps. In this study, we extend this approach to a dynamic, multi-agent simulation framework involving both cars and pedestrians. Starting from the initial fastest-path routing, we simulate the co-evolution of vehicular emissions and pedestrian exposure over multiple steps, where pedestrian flows dynamically influence car emissions, and vice versa. Two routing strategies are explored: global weighting, where a shared trade-off between travel time and exposure is selected, and local weighting, where each trip independently chooses its optimal trade-off. Experiments conducted on real-world urban data from a medium-sized city in Italy demonstrate that both strategies achieve significant reductions in pedestrian exposure; however, they differ in their impact on vehicle emissions and travel times. Global weighting yields more coordinated adaptation but at a higher systemic cost, while local weighting achieves more balanced outcomes with lower disruption. These results provide insights into designing urban routing policies that jointly optimize mobility efficiency and environmental sustainability.Source: MOBILE NETWORKS AND APPLICATIONS
DOI: 10.1007/s11036-025-02459-4
DOI: 10.21203/rs.3.rs-6603263/v1
Metrics:
Aliyev G., Nanni M.
Vehicular traffic is a major source of air pollution in urban areas, exposing pedestrians and residents to harmful emissions. Recent works have proposed exposure-aware pedestrian routing strategies based on static emission maps. In this study, we extend this approach to a dynamic, multi-agent simulation framework involving both cars and pedestrians. Starting from the initial fastest-path routing, we simulate the co-evolution of vehicular emissions and pedestrian exposure over multiple steps, where pedestrian flows dynamically influence car emissions, and vice versa. Two routing strategies are explored: global weighting, where a shared trade-off between travel time and exposure is selected, and local weighting, where each trip independently chooses its optimal trade-off. Experiments conducted on real-world urban data from a medium-sized city in Italy demonstrate that both strategies achieve significant reductions in pedestrian exposure; however, they differ in their impact on vehicle emissions and travel times. Global weighting yields more coordinated adaptation but at a higher systemic cost, while local weighting achieves more balanced outcomes with lower disruption. These results provide insights into designing urban routing policies that jointly optimize mobility efficiency and environmental sustainability.Source: MOBILE NETWORKS AND APPLICATIONS
DOI: 10.1007/s11036-025-02459-4
DOI: 10.21203/rs.3.rs-6603263/v1
Metrics:
See at:
doi.org | CNR IRIS | link.springer.com | Mobile Networks and Applications | CNR IRIS | CNR IRIS
2025
Contribution to book
Open Access
From GPS traces to individual emission exposure: a data-driven four-step process
Aliyev G., Nanni M.
Vehicular traffic is one of the major sources of air pollution in urban settings, making it essential to clearly understand how much and where vehicle emissions impact residents. Estimating vehicular pollution using GPS trajectories and microscopic models is getting more popular as this method has several advantages compared to other approaches. However, GPS data sources usually cover only a small sample of actual traffic, making current approaches unable to provide emission estimates for the whole road network. Moreover, to understand how much of these emissions reach different locations, a dispersion model should be applied, and quantifying their effect on individuals requires considering where they stay and/or how they move. Therefore, in this paper, we propose a four-step process that elaborates on raw, incomplete emission estimates and (i) first, estimates initial emissions from GPS data, (ii) estimates emission concentrations for the missing road segments, (iii) further processes the emission data to consider air dispersion, and (iv) computes the expected exposure to emissions of individuals in several use cases, involving both public buildings (e.g. schools) and pedestrian mobility. The experiments are based on a sample of vehicular GPS data in two Italian cities.Source: LECTURE NOTES OF THE INSTITUTE FOR COMPUTER SCIENCES, SOCIAL INFORMATICS AND TELECOMMUNICATIONS ENGINEERING (INTERNET), vol. 608, pp. 64-82
DOI: 10.1007/978-3-031-86370-7_5
Metrics:
Aliyev G., Nanni M.
Vehicular traffic is one of the major sources of air pollution in urban settings, making it essential to clearly understand how much and where vehicle emissions impact residents. Estimating vehicular pollution using GPS trajectories and microscopic models is getting more popular as this method has several advantages compared to other approaches. However, GPS data sources usually cover only a small sample of actual traffic, making current approaches unable to provide emission estimates for the whole road network. Moreover, to understand how much of these emissions reach different locations, a dispersion model should be applied, and quantifying their effect on individuals requires considering where they stay and/or how they move. Therefore, in this paper, we propose a four-step process that elaborates on raw, incomplete emission estimates and (i) first, estimates initial emissions from GPS data, (ii) estimates emission concentrations for the missing road segments, (iii) further processes the emission data to consider air dispersion, and (iv) computes the expected exposure to emissions of individuals in several use cases, involving both public buildings (e.g. schools) and pedestrian mobility. The experiments are based on a sample of vehicular GPS data in two Italian cities.Source: LECTURE NOTES OF THE INSTITUTE FOR COMPUTER SCIENCES, SOCIAL INFORMATICS AND TELECOMMUNICATIONS ENGINEERING (INTERNET), vol. 608, pp. 64-82
DOI: 10.1007/978-3-031-86370-7_5
Metrics:
See at:
CNR IRIS | link.springer.com | CNR IRIS | CNR IRIS
2025
Book
Open Access
Preface to INTSYS 2024
Kocian A., Milazzo P., Martins A. L., Nanni M., Pappalardo L.
An abstract is not available
Kocian A., Milazzo P., Martins A. L., Nanni M., Pappalardo L.
An abstract is not available
See at:
CNR IRIS | link.springer.com | CNR IRIS
2025
Contribution to book
Open Access
Preface to 7th International Workshop on Big Mobility Data Analytics (BMDA)
Nanni M., Pelekis N., Tampakis P., Zeitouni K., Sakr M., Renso C., Soares A., Artikis A., Theodoridis Y., Damiani M. L., Zissis D., Raffaetà A., Doulkeridis C., Kim K. -S., Ferhatosmanoglu H., Patroumpas K., Zeinalipour D., Coelho Da Silva T. L., Tserpes K., Andersen N. S., Pfoser D., Pappalardo L., Guidotti R., Kontopoulos I., Lu H., Nørvåg K., Andrienko N.
An abstract is not available
Nanni M., Pelekis N., Tampakis P., Zeitouni K., Sakr M., Renso C., Soares A., Artikis A., Theodoridis Y., Damiani M. L., Zissis D., Raffaetà A., Doulkeridis C., Kim K. -S., Ferhatosmanoglu H., Patroumpas K., Zeinalipour D., Coelho Da Silva T. L., Tserpes K., Andersen N. S., Pfoser D., Pappalardo L., Guidotti R., Kontopoulos I., Lu H., Nørvåg K., Andrienko N.
An abstract is not available
See at:
ceur-ws.org | CNR IRIS | CNR IRIS
2025
Conference article
Open Access
Explaining urban vehicle emissions in Rome
Bohm M., Reyes P., Nanni M., Pappalardo L.
Urban emissions are a significant challenge for city livability. Our work focuses on studying vehicle emissions in cities, using spatial and non-spatial models to understand their relationships with various urban features. We find that the spatial model demonstrates better performance and provides powerful insights into the influence of different predictors in various city areas. Our findings reveal that CO2 emissions in Rome are primarily linked to the presence of main arterial roads, population density, and road network density. However, the importance of these factors varies across different areas of the city. We also performed a what-if analysis to show that limiting the circulation of highly polluting vehicles may help reduce emissions, especially in city centres. Our research contributes to a better understanding of the complex relationships between the urban environment and the spatial variability of vehicle emissions in Rome.Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 15244, pp. 303-315. Pisa, Italy, 14–16/10/2024
DOI: 10.1007/978-3-031-78980-9_19
Project(s): SoBigData-PlusPlus
Metrics:
Bohm M., Reyes P., Nanni M., Pappalardo L.
Urban emissions are a significant challenge for city livability. Our work focuses on studying vehicle emissions in cities, using spatial and non-spatial models to understand their relationships with various urban features. We find that the spatial model demonstrates better performance and provides powerful insights into the influence of different predictors in various city areas. Our findings reveal that CO2 emissions in Rome are primarily linked to the presence of main arterial roads, population density, and road network density. However, the importance of these factors varies across different areas of the city. We also performed a what-if analysis to show that limiting the circulation of highly polluting vehicles may help reduce emissions, especially in city centres. Our research contributes to a better understanding of the complex relationships between the urban environment and the spatial variability of vehicle emissions in Rome.Source: LECTURE NOTES IN COMPUTER SCIENCE, vol. 15244, pp. 303-315. Pisa, Italy, 14–16/10/2024
DOI: 10.1007/978-3-031-78980-9_19
Project(s): SoBigData-PlusPlus
Metrics:
See at:
CNR IRIS | link.springer.com | doi.org | CNR IRIS | CNR IRIS
2025
Journal article
Open Access
Modeling Events and Interactions through Temporal Processes - A Survey
Liguori A., Caroprese L., Minici M., Veloso B., Spinnato F., Nanni M., Manco G., Gama J.
In real-world scenario, many phenomena produce a collection of events that occur in continuous time. Point Processes provide a natural mathematical framework for modeling these sequences of events. In this survey, we investigate probabilistic models for modeling event sequences through temporal processes. We revise the notion of event modeling and provide the mathematical foundations that characterize the literature on the topic. We define an ontology to categorize the existing approaches in terms of three families: simple, marked, and spatio-temporal point processes. For each family, we systematically review the existing approaches based based on deep learning. Finally, we analyze the scenarios where the proposed techniques can be used for addressing prediction and modeling aspects.Source: NEUROCOMPUTING, vol. 653 (issue 131191)
DOI: 10.1016/j.neucom.2025.131191
DOI: 10.48550/arxiv.2303.06067
Project(s): Automatic Retinal Macular Hole and Edema Diagnostics System, HumanE-AI-Net
Metrics:
Liguori A., Caroprese L., Minici M., Veloso B., Spinnato F., Nanni M., Manco G., Gama J.
In real-world scenario, many phenomena produce a collection of events that occur in continuous time. Point Processes provide a natural mathematical framework for modeling these sequences of events. In this survey, we investigate probabilistic models for modeling event sequences through temporal processes. We revise the notion of event modeling and provide the mathematical foundations that characterize the literature on the topic. We define an ontology to categorize the existing approaches in terms of three families: simple, marked, and spatio-temporal point processes. For each family, we systematically review the existing approaches based based on deep learning. Finally, we analyze the scenarios where the proposed techniques can be used for addressing prediction and modeling aspects.Source: NEUROCOMPUTING, vol. 653 (issue 131191)
DOI: 10.1016/j.neucom.2025.131191
DOI: 10.48550/arxiv.2303.06067
Project(s): Automatic Retinal Macular Hole and Edema Diagnostics System
, HumanE-AI-Net Metrics:
See at:
arXiv.org e-Print Archive | Neurocomputing | CNR IRIS | www.sciencedirect.com | Software Heritage | Software Heritage | IRIS Cnr | doi.org | GitHub | GitHub | GitHub | GitHub | GitHub | GitHub | GitHub | GitHub | IRIS Cnr | CNR IRIS
2025
Conference article
Open Access
Exploiting vehicular data for exposure-aware pedestrian routing
Aliyev G., Nanni M.
Vehicular traffic is one of the major sources of air pollution in urban settings, making it essential to clearly understand how much and where vehicle emissions impact residents. Recent approaches manage to yield pollution maps at the microscopic level by processing GPS trajectories of vehicles. That is achieved by applying mathematical models to estimate instantaneous emissions from GPS data, extending estimates to areas without data through missing data imputation, and further considering air dispersion factors. In this work, we leverage such inferred knowledge to implement an emission-aware pedestrian routing strategy and to study its impact on the reduction of exposure to vehicular pollutants and walking time. The study is realized through simulations of large masses of pedestrians over a medium-sized city in Italy, analyzing the interplay between the two factors - exposure versus walking time - in terms of time efficiency of paths and changes over existing habits both at a global and at an individual level. Experiments suggest that exposure-aware routing can yield a significant margin of improvement in health over most paths with minor effects on mobility, making it feasible and effective.DOI: 10.1109/mdm65600.2025.00022
Metrics:
Aliyev G., Nanni M.
Vehicular traffic is one of the major sources of air pollution in urban settings, making it essential to clearly understand how much and where vehicle emissions impact residents. Recent approaches manage to yield pollution maps at the microscopic level by processing GPS trajectories of vehicles. That is achieved by applying mathematical models to estimate instantaneous emissions from GPS data, extending estimates to areas without data through missing data imputation, and further considering air dispersion factors. In this work, we leverage such inferred knowledge to implement an emission-aware pedestrian routing strategy and to study its impact on the reduction of exposure to vehicular pollutants and walking time. The study is realized through simulations of large masses of pedestrians over a medium-sized city in Italy, analyzing the interplay between the two factors - exposure versus walking time - in terms of time efficiency of paths and changes over existing habits both at a global and at an individual level. Experiments suggest that exposure-aware routing can yield a significant margin of improvement in health over most paths with minor effects on mobility, making it feasible and effective.DOI: 10.1109/mdm65600.2025.00022
Metrics:
See at:
CNR IRIS | ieeexplore.ieee.org | doi.org | CNR IRIS | CNR IRIS
2024
Journal article
Open Access
Navigation services and urban sustainability
Cornacchia G, Nanni M, Pedreschi D, Pappalardo L
The The rise of socio-technical systems in which humans interact with various forms of Artificial Intelligence, including assistants and recommenders, multiplies the possibility for the emergence of large-scale social behavior, possibly with unintended negative consequences. In this work, we discuss a particularly interesting case, i.e., navigation services' impact on urban emissions, showing through simulations that the sum of many individually "optimal" choices may have unintended negative outcomes because such choices influence and interfere with each other on top of shared resources. To prove this point, we demonstrate how the introduction of a random component in the path suggestion phase may help to relieve the effect of collective and individual choices on the urban environment in terms of urban emissions.Source: FLUCTUATION AND NOISE LETTERS, vol. 23 (issue 3)
DOI: 10.1142/s0219477524500160
Project(s): HumanE-AI-Net, XAI , SoBigData-PlusPlus
Metrics:
Cornacchia G, Nanni M, Pedreschi D, Pappalardo L
The The rise of socio-technical systems in which humans interact with various forms of Artificial Intelligence, including assistants and recommenders, multiplies the possibility for the emergence of large-scale social behavior, possibly with unintended negative consequences. In this work, we discuss a particularly interesting case, i.e., navigation services' impact on urban emissions, showing through simulations that the sum of many individually "optimal" choices may have unintended negative outcomes because such choices influence and interfere with each other on top of shared resources. To prove this point, we demonstrate how the introduction of a random component in the path suggestion phase may help to relieve the effect of collective and individual choices on the urban environment in terms of urban emissions.Source: FLUCTUATION AND NOISE LETTERS, vol. 23 (issue 3)
DOI: 10.1142/s0219477524500160
Project(s): HumanE-AI-Net
, XAI , SoBigData-PlusPlus Metrics:
See at:
CNR IRIS | Fluctuation and Noise Letters | CNR IRIS | CNR IRIS
2024
Journal article
Open Access
Fast, interpretable, and deterministic time series classification with a bag-of-receptive-fields
Spinnato F., Guidotti R., Monreale A., Nanni M.
The current trend in the literature on Time Series Classification is to develop increasingly accurate algorithms by combining multiple models in ensemble hybrids, representing time series in complex and expressive feature spaces, and extracting features from different representations of the same time series. As a consequence of this focus on predictive performance, the best time series classifiers are black-box models, which are not understandable from a human standpoint. Even the approaches that are regarded as interpretable, such as shapelet-based ones, rely on randomization to maintain computational efficiency. This poses challenges for interpretability, as the explanation can change from run to run. Given these limitations, we propose the Bag-Of-Receptive-Field (BORF), a fast, interpretable, and deterministic time series transform. Building upon the classical Bag-Of-Patterns, we bridge the gap between convolutional operators and discretization, enhancing the Symbolic Aggregate Approximation (SAX) with dilation and stride, which can more effectively capture temporal patterns at multiple scales. We propose an algorithmic speedup that reduces the time complexity associated with SAX-based classifiers, allowing the extension of the Bag-Of-Patterns to the more flexible Bag-Of-Receptive-Fields, represented as a sparse multivariate tensor. The empirical results from testing our proposal on more than 150 univariate and multivariate classification datasets demonstrate good accuracy and great computational efficiency compared to traditional SAX-based methods and state-of-the-art time series classifiers, while providing easy-to-understand explanations.Source: IEEE ACCESS, vol. 12, pp. 137893-137912
DOI: 10.1109/access.2024.3464743
Project(s): Green.Dat.AI, TANGO , XAI , SoBigData-PlusPlus
Metrics:
Spinnato F., Guidotti R., Monreale A., Nanni M.
The current trend in the literature on Time Series Classification is to develop increasingly accurate algorithms by combining multiple models in ensemble hybrids, representing time series in complex and expressive feature spaces, and extracting features from different representations of the same time series. As a consequence of this focus on predictive performance, the best time series classifiers are black-box models, which are not understandable from a human standpoint. Even the approaches that are regarded as interpretable, such as shapelet-based ones, rely on randomization to maintain computational efficiency. This poses challenges for interpretability, as the explanation can change from run to run. Given these limitations, we propose the Bag-Of-Receptive-Field (BORF), a fast, interpretable, and deterministic time series transform. Building upon the classical Bag-Of-Patterns, we bridge the gap between convolutional operators and discretization, enhancing the Symbolic Aggregate Approximation (SAX) with dilation and stride, which can more effectively capture temporal patterns at multiple scales. We propose an algorithmic speedup that reduces the time complexity associated with SAX-based classifiers, allowing the extension of the Bag-Of-Patterns to the more flexible Bag-Of-Receptive-Fields, represented as a sparse multivariate tensor. The empirical results from testing our proposal on more than 150 univariate and multivariate classification datasets demonstrate good accuracy and great computational efficiency compared to traditional SAX-based methods and state-of-the-art time series classifiers, while providing easy-to-understand explanations.Source: IEEE ACCESS, vol. 12, pp. 137893-137912
DOI: 10.1109/access.2024.3464743
Project(s): Green.Dat.AI
, TANGO , XAI , SoBigData-PlusPlus Metrics:
See at:
IEEE Access | IEEE Access | CNR IRIS | ieeexplore.ieee.org | IRIS Cnr | Archivio della Ricerca - Università di Pisa | IRIS Cnr | CNR IRIS
2024
Conference article
Open Access
Barcelona effect: studying the instability of shortest paths in urban settings
Cornacchia G., Nanni M., Grassi F.
Human mobility is one of the important factors affecting the efficiency of cities and the quality of life of their dwellers. However, while city planners aim to improve the urban road network design to satisfy the local mobility demand and distribute traffic in an optimal way, the structure of cities across different areas and countries vary considerably and in complex ways, sometimes being the result of historical stratifications. One question that emerges, then, is how we can characterize cities in terms of (potential) traffic efficiency. In this work we aim to study the problem from a new perspective, introducing the concept of (shortest) path instability, which quantifies the tendency of a road network to provide very different travel alternatives for just slightly different trips. A notable case of that, which stimulated this research, is the city of Barcelona, where, apparently, reaching very close destinations might require very different routes. The concept is implemented and applied to two case studies at different spatial scales, one comparing the European capitals and the other comparing municipalities of an Italian region. Results show that path instability is heterogeneously distributed, with some largely unstable cities and others very stable, and it is not directly determined by simple city characteristics, such as the city size or its "smartness".Source: CEUR WORKSHOP PROCEEDINGS, vol. 3651. Paestum, Italy, 25/03/2025
Project(s): Green.Dat.AI
Cornacchia G., Nanni M., Grassi F.
Human mobility is one of the important factors affecting the efficiency of cities and the quality of life of their dwellers. However, while city planners aim to improve the urban road network design to satisfy the local mobility demand and distribute traffic in an optimal way, the structure of cities across different areas and countries vary considerably and in complex ways, sometimes being the result of historical stratifications. One question that emerges, then, is how we can characterize cities in terms of (potential) traffic efficiency. In this work we aim to study the problem from a new perspective, introducing the concept of (shortest) path instability, which quantifies the tendency of a road network to provide very different travel alternatives for just slightly different trips. A notable case of that, which stimulated this research, is the city of Barcelona, where, apparently, reaching very close destinations might require very different routes. The concept is implemented and applied to two case studies at different spatial scales, one comparing the European capitals and the other comparing municipalities of an Italian region. Results show that path instability is heterogeneously distributed, with some largely unstable cities and others very stable, and it is not directly determined by simple city characteristics, such as the city size or its "smartness".Source: CEUR WORKSHOP PROCEEDINGS, vol. 3651. Paestum, Italy, 25/03/2025
Project(s): Green.Dat.AI
See at:
ceur-ws.org | CNR IRIS | CNR IRIS
2024
Contribution to book
Open Access
Message from the MAURO 2024 Workshop Chairs
Chondrodima E., Cornacchia G., Mauro G., Nanni M., Pappalardo L., Pugliese C.
An abstract is not availableDOI: 10.1109/mdm61037.2024.00011
Metrics:
Chondrodima E., Cornacchia G., Mauro G., Nanni M., Pappalardo L., Pugliese C.
An abstract is not availableDOI: 10.1109/mdm61037.2024.00011
Metrics:
See at:
CNR IRIS | www.computer.org | doi.org | CNR IRIS
2024
Conference article
Open Access
On the pursuit of graph embedding strategies for individual mobility networks
Alamdari O. I., Nanni M., Bonavita A.
An Individual Mobility Network (IMN) is a graph representation of the mobility history of an individual that highlights the relevant locations visited (nodes of the graph) and the movements across them (edges), also providing a rich set of annotations of both nodes and edges. Extracting representative features from an IMN has proven to be a valuable task for enabling various learning applications. However, it is also a demanding operation that does not guarantee the inclusion of all important aspects from the human perspective. A vast recent literature on graph embedding goes in a similar direction, yet typically aims at general-purpose methods that might not suit specific contexts. In this paper, we discuss the existing approaches to graph embedding and the specificities of IMNs, trying to find the best matching solutions. We experiment with representative algorithms and study the results in relation to IMN characteristics. Tests are performed on a large dataset of real vehicle trajectories.DOI: 10.1109/bigdata59044.2023.10386373
Metrics:
Alamdari O. I., Nanni M., Bonavita A.
An Individual Mobility Network (IMN) is a graph representation of the mobility history of an individual that highlights the relevant locations visited (nodes of the graph) and the movements across them (edges), also providing a rich set of annotations of both nodes and edges. Extracting representative features from an IMN has proven to be a valuable task for enabling various learning applications. However, it is also a demanding operation that does not guarantee the inclusion of all important aspects from the human perspective. A vast recent literature on graph embedding goes in a similar direction, yet typically aims at general-purpose methods that might not suit specific contexts. In this paper, we discuss the existing approaches to graph embedding and the specificities of IMNs, trying to find the best matching solutions. We experiment with representative algorithms and study the results in relation to IMN characteristics. Tests are performed on a large dataset of real vehicle trajectories.DOI: 10.1109/bigdata59044.2023.10386373
Metrics:
See at:
Archivio istituzionale della Ricerca - Scuola Normale Superiore | Archivio istituzionale della Ricerca - Scuola Normale Superiore | CNR IRIS | ieeexplore.ieee.org | doi.org | CNR IRIS | CNR IRIS
2024
Other
Open Access
Quantifying and mitigating the impact of vehicular routing on the urban environment
Cornacchia G., Pappalardo L., Nanni Mirco
Urbanization pressures cities to efficiently accommodate the increasing demand for mobility, making traffic optimization challenging due to the complex interplay be- tween road networks and traffic dynamics, as drivers’ routing choices significantly in- fluence one another. City-related services, such as navigation services (e.g., TomTom) and mobility policies (e.g., road closures), impact traffic patterns and emissions. Nav- igation services can unintentionally increase emissions when many vehicles converge on the same routes, while mobility policies may have counterintuitive effects on traffic. We propose a simulation framework to assess the impact of road closure policies and navigation services on the urban environment. We use this framework and find that targeted road closures in Milan can reduce emissions by up to 10%, while others can increase emissions by nearly 50%. Then, we examine navigation services’ impact on vehicular traffic and CO2 emissions, finding that they reduce emissions at low traffic loads. However, at high traffic loads and penetration rates, they cause conformist behavior, leading to inefficiencies and potentially higher emissions. To mitigate the conformist behavior induced by navigation services and reduce CO2 emissions, we propose three solutions: (i) an individualistic approach using existing Alternative Routing (AR) algorithms, (ii) Metis, a coordinated solution that coordinates drivers and dynamically estimates traffic to diversify routes, and (iii) Polaris, an individual AR algorithm which considers road popularity to optimize traffic distribution. Moti- vated by the varying effectiveness of AR solutions across cities, we study cities’ route diversification, defining shortest path instability and introducing diverCity, a metric to assess a city’s propensity towards route diversity. Analysis shows that diverCity benefits from extensive road networks, leading to less congestion. We also address the impact of mobility attractors on diverCity and propose mitigation strategies. This thesis comprehensively studies vehicular traffic dynamics, offering a simulation framework to evaluate the environmental impact of mobility policies and navigation services. In addition, it presents solutions to mitigate negative impacts and proposes metrics to quantify a city’s potential to offer route diversity.
Cornacchia G., Pappalardo L., Nanni Mirco
Urbanization pressures cities to efficiently accommodate the increasing demand for mobility, making traffic optimization challenging due to the complex interplay be- tween road networks and traffic dynamics, as drivers’ routing choices significantly in- fluence one another. City-related services, such as navigation services (e.g., TomTom) and mobility policies (e.g., road closures), impact traffic patterns and emissions. Nav- igation services can unintentionally increase emissions when many vehicles converge on the same routes, while mobility policies may have counterintuitive effects on traffic. We propose a simulation framework to assess the impact of road closure policies and navigation services on the urban environment. We use this framework and find that targeted road closures in Milan can reduce emissions by up to 10%, while others can increase emissions by nearly 50%. Then, we examine navigation services’ impact on vehicular traffic and CO2 emissions, finding that they reduce emissions at low traffic loads. However, at high traffic loads and penetration rates, they cause conformist behavior, leading to inefficiencies and potentially higher emissions. To mitigate the conformist behavior induced by navigation services and reduce CO2 emissions, we propose three solutions: (i) an individualistic approach using existing Alternative Routing (AR) algorithms, (ii) Metis, a coordinated solution that coordinates drivers and dynamically estimates traffic to diversify routes, and (iii) Polaris, an individual AR algorithm which considers road popularity to optimize traffic distribution. Moti- vated by the varying effectiveness of AR solutions across cities, we study cities’ route diversification, defining shortest path instability and introducing diverCity, a metric to assess a city’s propensity towards route diversity. Analysis shows that diverCity benefits from extensive road networks, leading to less congestion. We also address the impact of mobility attractors on diverCity and propose mitigation strategies. This thesis comprehensively studies vehicular traffic dynamics, offering a simulation framework to evaluate the environmental impact of mobility policies and navigation services. In addition, it presents solutions to mitigate negative impacts and proposes metrics to quantify a city’s potential to offer route diversity.
2023
Conference article
Open Access
Geolet: an interpretable model for trajectory classification
Landi C, Spinnato F, Guidotti R, Monreale A, Nanni M
The large and diverse availability of mobility data enables the development of predictive models capable of recognizing various types of movements. Through a variety of GPS devices, any moving entity, animal, person, or vehicle can generate spatio-temporal trajectories. This data is used to infer migration patterns, manage traffic in large cities, and monitor the spread and impact of diseases, all critical situations that necessitate a thorough understanding of the underlying problem. Researchers, businesses, and governments use mobility data to make decisions that affect people's lives in many ways, employing accurate but opaque deep learning models that are difficult to interpret from a human standpoint. To address these limitations, we propose Geolet, a human-interpretable machine-learning model for trajectory classification. We use discriminative sub-trajectories extracted from mobility data to turn trajectories into a simplified representation that can be used as input by any machine learning classifier. We test our approach against state-of-the-art competitors on real-world datasets. Geolet outperforms black-box models in terms of accuracy while being orders of magnitude faster than its interpretable competitors.DOI: 10.1007/978-3-031-30047-9_19
Project(s): TAILOR, XAI , SoBigData-PlusPlus , Humane AI
Metrics:
Landi C, Spinnato F, Guidotti R, Monreale A, Nanni M
The large and diverse availability of mobility data enables the development of predictive models capable of recognizing various types of movements. Through a variety of GPS devices, any moving entity, animal, person, or vehicle can generate spatio-temporal trajectories. This data is used to infer migration patterns, manage traffic in large cities, and monitor the spread and impact of diseases, all critical situations that necessitate a thorough understanding of the underlying problem. Researchers, businesses, and governments use mobility data to make decisions that affect people's lives in many ways, employing accurate but opaque deep learning models that are difficult to interpret from a human standpoint. To address these limitations, we propose Geolet, a human-interpretable machine-learning model for trajectory classification. We use discriminative sub-trajectories extracted from mobility data to turn trajectories into a simplified representation that can be used as input by any machine learning classifier. We test our approach against state-of-the-art competitors on real-world datasets. Geolet outperforms black-box models in terms of accuracy while being orders of magnitude faster than its interpretable competitors.DOI: 10.1007/978-3-031-30047-9_19
Project(s): TAILOR
, XAI , SoBigData-PlusPlus , Humane AI Metrics:
See at:
CNR IRIS | link.springer.com | ISTI Repository | doi.org | CNR IRIS | CNR IRIS
2023
Conference article
Open Access
The effects of route randomization on urban emissions
Cornacchia G, Nanni M, Pedreschi D, Pappalardo L
Routing algorithms typically suggest the fastest path or slight variation to reach a user's desired destination. Although this suggestion at the individual level is undoubtedly advantageous for the user, from a collective point of view, the aggregation of all single suggested paths may result in an increasing impact (e.g., in terms of emissions).In this study, we use SUMO to simulate the effects of incorporating randomness into routing algorithms on emissions, their distribution, and travel time in the urban area of Milan (Italy). Our results reveal that, given the common practice of routing towards the fastest path, a certain level of randomness in routes reduces emissions and travel time. In other words, the stronger the random component in the routes, the more pronounced the benefits upon a certain threshold. Our research provides insight into the potential advantages of considering collective outcomes in routing decisions and highlights the need to explore further the relationship between route randomization and sustainability in urban transportation.DOI: 10.52825/scp.v4i.217
Project(s): SoBigData-PlusPlus
Metrics:
Cornacchia G, Nanni M, Pedreschi D, Pappalardo L
Routing algorithms typically suggest the fastest path or slight variation to reach a user's desired destination. Although this suggestion at the individual level is undoubtedly advantageous for the user, from a collective point of view, the aggregation of all single suggested paths may result in an increasing impact (e.g., in terms of emissions).In this study, we use SUMO to simulate the effects of incorporating randomness into routing algorithms on emissions, their distribution, and travel time in the urban area of Milan (Italy). Our results reveal that, given the common practice of routing towards the fastest path, a certain level of randomness in routes reduces emissions and travel time. In other words, the stronger the random component in the routes, the more pronounced the benefits upon a certain threshold. Our research provides insight into the potential advantages of considering collective outcomes in routing decisions and highlights the need to explore further the relationship between route randomization and sustainability in urban transportation.DOI: 10.52825/scp.v4i.217
Project(s): SoBigData-PlusPlus
Metrics:
See at:
CNR IRIS | ISTI Repository | www.tib-op.org | CNR IRIS
2023
Journal article
Open Access
Understanding any time series classifier with a subsequence-based explainer
Spinnato F, Guidotti R, Monreale A, Nanni M, Pedreschi D, Giannotti F
The growing availability of time series data has increased the usage of classifiers for this data type. Unfortunately, state-of-the-art time series classifiers are black-box models and, therefore, not usable in critical domains such as healthcare or finance, where explainability can be a crucial requirement. This paper presents a framework to explain the predictions of any black-box classifier for univariate and multivariate time series. The provided explanation is composed of three parts. First, a saliency map highlighting the most important parts of the time series for the classification. Second, an instance-based explanation exemplifies the blackbox's decision by providing a set of prototypical and counterfactual time series. Third, a factual and counterfactual rule-based explanation, revealing the reasons for the classification through logical conditions based on subsequences that must, or must not, be contained in the time series. Experiments and benchmarks show that the proposed method provides faithful, meaningful, stable, and interpretable explanations.Source: ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA, vol. 18 (issue 2), pp. 1-34
DOI: 10.1145/3624480
Project(s): TAILOR, XAI , SoBigData-PlusPlus
Metrics:
Spinnato F, Guidotti R, Monreale A, Nanni M, Pedreschi D, Giannotti F
The growing availability of time series data has increased the usage of classifiers for this data type. Unfortunately, state-of-the-art time series classifiers are black-box models and, therefore, not usable in critical domains such as healthcare or finance, where explainability can be a crucial requirement. This paper presents a framework to explain the predictions of any black-box classifier for univariate and multivariate time series. The provided explanation is composed of three parts. First, a saliency map highlighting the most important parts of the time series for the classification. Second, an instance-based explanation exemplifies the blackbox's decision by providing a set of prototypical and counterfactual time series. Third, a factual and counterfactual rule-based explanation, revealing the reasons for the classification through logical conditions based on subsequences that must, or must not, be contained in the time series. Experiments and benchmarks show that the proposed method provides faithful, meaningful, stable, and interpretable explanations.Source: ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA, vol. 18 (issue 2), pp. 1-34
DOI: 10.1145/3624480
Project(s): TAILOR
, XAI , SoBigData-PlusPlus Metrics:
See at:
dl.acm.org | CNR IRIS | ISTI Repository | ACM Transactions on Knowledge Discovery from Data | CNR IRIS
2023
Conference article
Open Access
Human mobility, AI assistants, and urban emissions: an insidious triangle
Pappalardo L, Bohm M, Cornacchia G, Mauro G, Pedreschi D, Nanni M
Transportation remains a significant contributor to greenhouse gas emissions, with a substantial proportion originating from road transport and passenger travel in particular. Today, the relationship between transportation and urban emissions is even more complex, given the increasingly prevalent role and the pervasiveness of AI-based GPS navigation systems such as Google Maps and TomTom. While these services offer benefits to individual drivers, they can also exacerbate congestion and increase pollution if too many drivers are directed onto the same route. In this article, we provide two examples from our research group that explore the impact of vehicular transportation and mobility-AI-based applications on urban emissions. By conducting realistic simulations and studying the impact of GPS navigation systems on emissions, we provide insights into the potential for mitigating transportation emissions and developing policies that promote sustainable urban mobility. Our examples demonstrate how vehicle-generated emissions can be reduced and how studying the impact of GPS navigation systems on emissions can lead to unexpected findings. Overall, our analysis suggests that it is crucial to consider the impact of emerging technologies on transportation and emissions, and to develop strategies that promote sustainable mobility while ensuring the optimal use of these tools.Source: CEUR WORKSHOP PROCEEDINGS, pp. 585-589. Pisa, Italy, 29-31/05/2023
Project(s): HumanE-AI-Net, SoBigData-PlusPlus
Pappalardo L, Bohm M, Cornacchia G, Mauro G, Pedreschi D, Nanni M
Transportation remains a significant contributor to greenhouse gas emissions, with a substantial proportion originating from road transport and passenger travel in particular. Today, the relationship between transportation and urban emissions is even more complex, given the increasingly prevalent role and the pervasiveness of AI-based GPS navigation systems such as Google Maps and TomTom. While these services offer benefits to individual drivers, they can also exacerbate congestion and increase pollution if too many drivers are directed onto the same route. In this article, we provide two examples from our research group that explore the impact of vehicular transportation and mobility-AI-based applications on urban emissions. By conducting realistic simulations and studying the impact of GPS navigation systems on emissions, we provide insights into the potential for mitigating transportation emissions and developing policies that promote sustainable urban mobility. Our examples demonstrate how vehicle-generated emissions can be reduced and how studying the impact of GPS navigation systems on emissions can lead to unexpected findings. Overall, our analysis suggests that it is crucial to consider the impact of emerging technologies on transportation and emissions, and to develop strategies that promote sustainable mobility while ensuring the optimal use of these tools.Source: CEUR WORKSHOP PROCEEDINGS, pp. 585-589. Pisa, Italy, 29-31/05/2023
Project(s): HumanE-AI-Net
, SoBigData-PlusPlus
See at:
ceur-ws.org | CNR IRIS | ISTI Repository | CNR IRIS
2023
Conference article
Restricted
Interpretable data partitioning through tree-based clustering methods
Guidotti R, Landi C, Beretta A, Fadda D, Nanni M
Interpretable Data Partitioning Through Tree-Based Clustering Methods Riccardo Guidotti, Cristiano Landi, Andrea Beretta, Daniele Fadda & Mirco Nanni Conference paper First Online: 08 October 2023 311 Accesses Part of the Lecture Notes in Computer Science book series (LNAI,volume 14276) The growing interpretable machine learning research field is mainly focusing on the explanation of supervised approaches. However, also unsupervised approaches might benefit from considering interpretability aspects. While existing clustering methods only provide the assignment of records to clusters without justifying the partitioning, we propose tree-based clustering methods that offer interpretable data partitioning through a shallow decision tree. These decision trees enable easy-to-understand explanations of cluster assignments through short and understandable split conditions. The proposed methods are evaluated through experiments on synthetic and real datasets and proved to be more effective than traditional clustering approaches and interpretable ones in terms of standard evaluation measures and runtime. Finally, a case study involving human participation demonstrates the effectiveness of the interpretable clustering trees returned by the proposed method.DOI: 10.1007/978-3-031-45275-8_33
Metrics:
Guidotti R, Landi C, Beretta A, Fadda D, Nanni M
Interpretable Data Partitioning Through Tree-Based Clustering Methods Riccardo Guidotti, Cristiano Landi, Andrea Beretta, Daniele Fadda & Mirco Nanni Conference paper First Online: 08 October 2023 311 Accesses Part of the Lecture Notes in Computer Science book series (LNAI,volume 14276) The growing interpretable machine learning research field is mainly focusing on the explanation of supervised approaches. However, also unsupervised approaches might benefit from considering interpretability aspects. While existing clustering methods only provide the assignment of records to clusters without justifying the partitioning, we propose tree-based clustering methods that offer interpretable data partitioning through a shallow decision tree. These decision trees enable easy-to-understand explanations of cluster assignments through short and understandable split conditions. The proposed methods are evaluated through experiments on synthetic and real datasets and proved to be more effective than traditional clustering approaches and interpretable ones in terms of standard evaluation measures and runtime. Finally, a case study involving human participation demonstrates the effectiveness of the interpretable clustering trees returned by the proposed method.DOI: 10.1007/978-3-031-45275-8_33
Metrics:
See at:
doi.org | CNR IRIS | CNR IRIS | link.springer.com
2023
Conference article
Open Access
TrajParquet: a trajectory-oriented column file format for mobility data lakes
Koutroumanis N, Doulkeridis C, Renso C, Nanni M, Perego R
Columnar data formats, such as Apache Parquet, are increasingly popular nowadays for scalable data storage and querying data lakes, due to compressed storage and efficient data access via data skipping. However, when applied to spatial or spatio-temporal data, advanced solutions are required to go beyond pruning over single attributes and towards multidimensional pruning. Even though there exist solutions for geospatial data, such as GeoParquet and SpatialParquet, they fall short when applied to trajectory data (sequences of spatio-temporal positions). In this paper, we propose TrajParquet, a format for columnar storage of trajectory data, which is highly efficient and scalable. Also, we present a query processing algorithm that supports spatio-temporal range queries over TrajParquet. We evaluate TrajParquet using real-world data sets and in comparison with extensions of GeoParquet and SpatialParquet, suitable for handling spatio-temporal data.DOI: 10.1145/3589132.3625623
Project(s): SoBigData-PlusPlus
Metrics:
Koutroumanis N, Doulkeridis C, Renso C, Nanni M, Perego R
Columnar data formats, such as Apache Parquet, are increasingly popular nowadays for scalable data storage and querying data lakes, due to compressed storage and efficient data access via data skipping. However, when applied to spatial or spatio-temporal data, advanced solutions are required to go beyond pruning over single attributes and towards multidimensional pruning. Even though there exist solutions for geospatial data, such as GeoParquet and SpatialParquet, they fall short when applied to trajectory data (sequences of spatio-temporal positions). In this paper, we propose TrajParquet, a format for columnar storage of trajectory data, which is highly efficient and scalable. Also, we present a query processing algorithm that supports spatio-temporal range queries over TrajParquet. We evaluate TrajParquet using real-world data sets and in comparison with extensions of GeoParquet and SpatialParquet, suitable for handling spatio-temporal data.DOI: 10.1145/3589132.3625623
Project(s): SoBigData-PlusPlus
Metrics:
See at:
CNR IRIS | ISTI Repository | CNR IRIS