Lucia Cavallaro
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lucia Cavallaro has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Antonio Liotta
Free University of Bozen-Bolzano
15 shared publications - Ovidiu Bagdasar
Data Science Research Centre, College of Science & Engineering, University of Derby, Derby DE22 1GB, UK
8 shared publications Pasquale De Meo
DICAM Department, University of Messina, Messina, Italy
7 shared publicationsGiacomo Fiumara
University of Messina, Department of Mathematics and Computer Science, Physics and Natural Science
5 shared publications- Giacomo Fiumara
MIFT Department, University of Messina, Messina, Italy
5 shared publications - Annamaria Ficara
University of Messina, Messina, Italy
4 shared publications
Latest Publications
- Open Access
Exploring the impact of adaptive rewiring in Graph Neural Networks
This paper explores sparsification methods as a form of regularization in Graph Neural Networks (GNNs) to address high memory usage and computational costs in large-scale graph applications. Using techniques from Network Science and Machine Learning, including Erdős-Rényi for model sparsification, we enhance the efficiency of GNNs for real-world applications. We demonstrate our approach on N-1 contingency assessment in electrical grids, a critical task for ensuring grid reliability. We apply our methods to three datasets of varying sizes, exploring Graph Convolutional Networks (GCN) and Graph Isomorphism Networks (GIN) with different degrees of sparsification and rewiring. Comparison across sparsification levels shows the potential of combining insights from both research fields to improve GNN performance and scalability. Our experiments highlight the importance of tuning sparsity parameters: while sparsity can improve generalization, excessive sparsity may hinder learning of complex patterns. Our adaptive rewiring approach, particularly when combined with early stopping, proves promising by allowing the model to adapt its connectivity structure during training. This research contributes to understanding how sparsity can be effectively leveraged in GNNs for critical applications like power grid reliability analysis.
Open AccessExploring the impact of adaptive rewiring in Graph Neural Networks
This paper explores sparsification methods as a form of regularization in Graph Neural Networks (GNNs) to address high memory usage and computational costs in large-scale graph applications. Using techniques from Network Science and Machine Learning, including Erdős-Rényi for model sparsification, we enhance the efficiency of GNNs for real-world applications. We demonstrate our approach on N-1 contingency assessment in electrical grids, a critical task for ensuring grid reliability. We apply our methods to three datasets of varying sizes, exploring Graph Convolutional Networks (GCN) and Graph Isomorphism Networks (GIN) with different degrees of sparsification and rewiring. Comparison across sparsification levels shows the potential of combining insights from both research fields to improve GNN performance and scalability. Our experiments highlight the importance of tuning sparsity parameters: while sparsity can improve generalization, excessive sparsity may hinder learning of complex patterns. Our adaptive rewiring approach, particularly when combined with early stopping, proves promising by allowing the model to adapt its connectivity structure during training. This research contributes to understanding how sparsity can be effectively leveraged in GNNs for critical applications like power grid reliability analysis.
- Open Access
- 5 October 2025
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 112
An experimental comparison of deep learning models for pneumonia classification from chest X-ray images
- 5 October 2025
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 112
Open AccessAn experimental comparison of deep learning models for pneumonia classification from chest X-ray images
- Open Access
Sparse Self-Federated Learning for Energy Efficient Cooperative Intelligence in Society 5.0
Federated Learning offers privacy-preserving collaborative intelligence but struggles to meet the sustainability demands of emerging IoT ecosystems necessary for Society 5.0-a human-centered technological future balancing social advancement with environmental responsibility. The excessive communication bandwidth and computational resources required by traditional FL approaches make them environmentally unsustainable at scale, creating a fundamental conflict with green AI principles as billions of resource-constrained devices attempt to participate. To this end, we introduce Sparse Proximity-based Self-Federated Learning (SParSeFuL), a resource-aware approach that bridges this gap by combining aggregate computing for self-organization with neural network sparsification to reduce energy and bandwidth consumption.
Open AccessSparse Self-Federated Learning for Energy Efficient Cooperative Intelligence in Society 5.0
Federated Learning offers privacy-preserving collaborative intelligence but struggles to meet the sustainability demands of emerging IoT ecosystems necessary for Society 5.0-a human-centered technological future balancing social advancement with environmental responsibility. The excessive communication bandwidth and computational resources required by traditional FL approaches make them environmentally unsustainable at scale, creating a fundamental conflict with green AI principles as billions of resource-constrained devices attempt to participate. To this end, we introduce Sparse Proximity-based Self-Federated Learning (SParSeFuL), a resource-aware approach that bridges this gap by combining aggregate computing for self-organization with neural network sparsification to reduce energy and bandwidth consumption.
- 3 June 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Modern Physics C
- Vol. 37 (06)
Adaptive fusion of multi-view learnable enhanced contrastive recommendation
Graph contrast learning effectively mitigates the problems of data sparsity and noise in the recommendation domain. However, current graph contrast learning algorithms still have certain shortcomings: (1) Most of them ignore the exploration and utilization of higher-order information, which may lead to the inability to make full use of the user information to obtain coarse-grained user representations; (2) Traditional graph contrast learning algorithms use random or manual augmentation to generate the contrast view, which may destroy the original topology of the interaction graph and amplify the effect of noise, which cannot produce an optimal contrast view. For this reason, an adaptive fusion multi-view learning enhanced contrast recommendation method (AMLECR) is proposed. First, a collaborative neighborhood graph is constructed based on the user-item interaction graph, and embeddings learned from higher-order information are utilized to achieve fine-grained user representations. A nonparametric adaptive fusion strategy is designed. Second, a multi-head attention network is introduced as an augmented view generator, which adaptively generates high-quality node-level augmented views. Finally, comparative analysis experiments are carried out on four real-world datasets such as Yelp with eight baselines. The experimental results show that the proposed AMLECR method has 1.4% and 2.7% improvement in Recall and NDCG metrics than the strongest baseline, which fully verifies the rationality and effectiveness of the proposed AMLECR method.
- 3 June 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Modern Physics C
- Vol. 37 (06)
Adaptive fusion of multi-view learnable enhanced contrastive recommendation
Graph contrast learning effectively mitigates the problems of data sparsity and noise in the recommendation domain. However, current graph contrast learning algorithms still have certain shortcomings: (1) Most of them ignore the exploration and utilization of higher-order information, which may lead to the inability to make full use of the user information to obtain coarse-grained user representations; (2) Traditional graph contrast learning algorithms use random or manual augmentation to generate the contrast view, which may destroy the original topology of the interaction graph and amplify the effect of noise, which cannot produce an optimal contrast view. For this reason, an adaptive fusion multi-view learning enhanced contrast recommendation method (AMLECR) is proposed. First, a collaborative neighborhood graph is constructed based on the user-item interaction graph, and embeddings learned from higher-order information are utilized to achieve fine-grained user representations. A nonparametric adaptive fusion strategy is designed. Second, a multi-head attention network is introduced as an augmented view generator, which adaptively generates high-quality node-level augmented views. Finally, comparative analysis experiments are carried out on four real-world datasets such as Yelp with eight baselines. The experimental results show that the proposed AMLECR method has 1.4% and 2.7% improvement in Recall and NDCG metrics than the strongest baseline, which fully verifies the rationality and effectiveness of the proposed AMLECR method.
- 11 April 2025
- Published by Springer Nature
- p. 3-14
A Comparative Study of Neural Network Pruning Strategies for AI Deployment on Edge Devices
- 11 April 2025
- Published by Springer Nature
- p. 3-14
A Comparative Study of Neural Network Pruning Strategies for AI Deployment on Edge Devices
- 1 January 2025
- Published by Springer Nature
- p. 296-303
Miniaturisation of Binary Classifiers Through Sparse Neural Networks
- 1 January 2025
- Published by Springer Nature
- p. 296-303
Miniaturisation of Binary Classifiers Through Sparse Neural Networks
- 4 December 2024
- Published by Springer Nature
- p. 150-159
Hybrid Edge-Cloud Federated Learning: The Case of Lightweight Smoking Detection
- 4 December 2024
- Published by Springer Nature
- p. 150-159
Hybrid Edge-Cloud Federated Learning: The Case of Lightweight Smoking Detection
- 28 November 2024
- Published by Springer Nature
- p. 101-106
Comparing Training of Sparse to Classic Neural Networks for Binary Classification in Medical Data
- 28 November 2024
- Published by Springer Nature
- p. 101-106
Comparing Training of Sparse to Classic Neural Networks for Binary Classification in Medical Data
- Open Access
- 9 May 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (5) , e0299255
On the sensitivity of centrality metrics
Despite the huge importance that the centrality metrics have in understanding the topology of a network, too little is known about the effects that small alterations in the topology of the input graph induce in the norm of the vector that stores the node centralities. If so, then it could be possible to avoid re-calculating the vector of centrality metrics if some minimal changes occur in the network topology, which would allow for significant computational savings. Hence, after formalising the notion of centrality, three of the most basic metrics were herein considered (i.e., Degree, Eigenvector, and Katz centrality). To perform the simulations, two probabilistic failure models were used to describe alterations in network topology: Uniform (i.e., all nodes can be independently deleted from the network with a fixed probability) and Best Connected (i.e., the probability a node is removed depends on its degree). Our analysis suggests that, in the case of degree, small variations in the topology of the input graph determine small variations in Degree centrality, independently of the topological features of the input graph; conversely, both Eigenvector and Katz centralities can be extremely sensitive to changes in the topology of the input graph. In other words, if the input graph has some specific features, even small changes in the topology of the input graph can have catastrophic effects on the Eigenvector or Katz centrality.
- 9 May 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (5) , e0299255
Open AccessOn the sensitivity of centrality metrics
Despite the huge importance that the centrality metrics have in understanding the topology of a network, too little is known about the effects that small alterations in the topology of the input graph induce in the norm of the vector that stores the node centralities. If so, then it could be possible to avoid re-calculating the vector of centrality metrics if some minimal changes occur in the network topology, which would allow for significant computational savings. Hence, after formalising the notion of centrality, three of the most basic metrics were herein considered (i.e., Degree, Eigenvector, and Katz centrality). To perform the simulations, two probabilistic failure models were used to describe alterations in network topology: Uniform (i.e., all nodes can be independently deleted from the network with a fixed probability) and Best Connected (i.e., the probability a node is removed depends on its degree). Our analysis suggests that, in the case of degree, small variations in the topology of the input graph determine small variations in Degree centrality, independently of the topological features of the input graph; conversely, both Eigenvector and Katz centralities can be extremely sensitive to changes in the topology of the input graph. In other words, if the input graph has some specific features, even small changes in the topology of the input graph can have catastrophic effects on the Eigenvector or Katz centrality.
- 20 February 2024
- Published by Springer Nature
- p. 331-338
Multi-class Classification Performance Improvements Through High Sparsity Strategies
- 20 February 2024
- Published by Springer Nature
- p. 331-338
Multi-class Classification Performance Improvements Through High Sparsity Strategies