Rodolfo Zunino
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Rodolfo Zunino has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
- Paolo Gastaldo
DITEN, University of Genoa, Genoa, Italy
27 shared publications - Edoardo Ragusa
Department of Electrical, Electronic, Telecommunications Engineering and Naval Architecture, University of Genoa, Genoa, Italy
16 shared publications - Christian Gianoglio
University of Genoa
7 shared publications - Federica Bisio
University of Genoa
5 shared publications - Sergio Decherchi
Fondazione Istituto Italiano di Tecnologia, Genoa, Italy
4 shared publications - Erik Cambria
Nanyang technological university, Singapore, Singapore
3 shared publications
Latest Publications
- 1 May 2026
- Published by Springer Nature
- p. 265-271
Sensing Parameters Versus Accuracy for Energy-Efficient Accelerometer Sensors
- 1 May 2026
- Published by Springer Nature
- p. 265-271
Sensing Parameters Versus Accuracy for Energy-Efficient Accelerometer Sensors
- Open Access
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 2982-2995
Hardware-Aware Neural Architecture Search for Encrypted Traffic Classification on Resource-Constrained Devices
This paper presents a hardware-efficient deep neural network (DNN), optimized through hardware-aware neural architecture search (HW-NAS); the DNN supports the classification of session-level encrypted traffic on resource-constrained Internet of Things (IoT) and edge devices. Thanks to HW-NAS, a 1D convolutional neural network (CNN) is tailored on the ISCX VPN-nonVPN dataset to meet strict memory and computational limits while achieving robust performance. The optimized model attains an accuracy of 96.60% with just 88.26K parameters, 10.08M floating-point operations (FLOPs), and a maximum tensor size of 20.12K. Compared to state-of-the-art (SOTA) models, it achieves reductions of up to 444-fold, 312-fold, and 15-fold in these metrics, respectively, significantly minimizing memory footprint and runtime requirements. The model also demonstrates versatility, achieving up to 99.86% across multiple VPN and traffic classification (TC) tasks; it further generalizes to external benchmarks with up to 99.98% accuracy on USTC-TFC and QUIC NetFlow. In addition, an in-depth approach to header-level preprocessing strategies confirms that the optimized model can provide notable performance across a wide range of configurations, even in scenarios with stricter privacy considerations. Likewise, a reduction in the length of sessions of up to 75% yields significant improvements in efficiency, while maintaining high accuracy with only a negligible drop of 1-2%. However, the importance of careful preprocessing and session length selection in the classification of raw traffic data is still present, as improper settings or aggressive reductions can bring about a 7% reduction in overall accuracy. The quantized architecture was deployed on STM32 microcontrollers and evaluated across input sizes; results confirm that the efficiency gains from shorter sessions translate to practical, low-latency embedded inference. These findings demonstrate the method’s practicality for encrypted traffic analysis in constrained IoT networks.
- 20 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 2982-2995
Open AccessHardware-Aware Neural Architecture Search for Encrypted Traffic Classification on Resource-Constrained Devices
This paper presents a hardware-efficient deep neural network (DNN), optimized through hardware-aware neural architecture search (HW-NAS); the DNN supports the classification of session-level encrypted traffic on resource-constrained Internet of Things (IoT) and edge devices. Thanks to HW-NAS, a 1D convolutional neural network (CNN) is tailored on the ISCX VPN-nonVPN dataset to meet strict memory and computational limits while achieving robust performance. The optimized model attains an accuracy of 96.60% with just 88.26K parameters, 10.08M floating-point operations (FLOPs), and a maximum tensor size of 20.12K. Compared to state-of-the-art (SOTA) models, it achieves reductions of up to 444-fold, 312-fold, and 15-fold in these metrics, respectively, significantly minimizing memory footprint and runtime requirements. The model also demonstrates versatility, achieving up to 99.86% across multiple VPN and traffic classification (TC) tasks; it further generalizes to external benchmarks with up to 99.98% accuracy on USTC-TFC and QUIC NetFlow. In addition, an in-depth approach to header-level preprocessing strategies confirms that the optimized model can provide notable performance across a wide range of configurations, even in scenarios with stricter privacy considerations. Likewise, a reduction in the length of sessions of up to 75% yields significant improvements in efficiency, while maintaining high accuracy with only a negligible drop of 1-2%. However, the importance of careful preprocessing and session length selection in the classification of raw traffic data is still present, as improper settings or aggressive reductions can bring about a 7% reduction in overall accuracy. The quantized architecture was deployed on STM32 microcontrollers and evaluated across input sizes; results confirm that the efficiency gains from shorter sessions translate to practical, low-latency embedded inference. These findings demonstrate the method’s practicality for encrypted traffic analysis in constrained IoT networks.
- Open Access
A Two-Dimensional (2-D) Sensor Network Architecture with Artificial Intelligence Models for the Detection of Magnetic Anomalies
The paper presents the development and preliminary evaluation of a two-dimensional (2-D) network of magnetometers for magnetic anomaly detection. The configuration significantly improves over the existing one-dimensional (1-D) architecture, as it enhances the spatial characterization of magnetic anomalies through the simultaneous acquisition of data over an extended area. This leads to a reliable estimation of the target motion parameters. Each sensor node in the network includes a custom-designed electronic system, integrating a biaxial fluxgate magnetometer that operates in null mode. Deep learning models process the raw measurements collected by the magnetometers and extract structured information that enables both automated detection and preliminary target tracking. In the experimental evaluation, a array of nodes was deployed over a m2 area for terrestrial tests, using moving ferromagnetic cylinders as targets. The results confirmed the feasibility of the 2-D configuration and supported its integration into intelligent, real-time surveillance systems for security and underwater monitoring applications.Open AccessA Two-Dimensional (2-D) Sensor Network Architecture with Artificial Intelligence Models for the Detection of Magnetic Anomalies
The paper presents the development and preliminary evaluation of a two-dimensional (2-D) network of magnetometers for magnetic anomaly detection. The configuration significantly improves over the existing one-dimensional (1-D) architecture, as it enhances the spatial characterization of magnetic anomalies through the simultaneous acquisition of data over an extended area. This leads to a reliable estimation of the target motion parameters. Each sensor node in the network includes a custom-designed electronic system, integrating a biaxial fluxgate magnetometer that operates in null mode. Deep learning models process the raw measurements collected by the magnetometers and extract structured information that enables both automated detection and preliminary target tracking. In the experimental evaluation, a array of nodes was deployed over a m2 area for terrestrial tests, using moving ferromagnetic cylinders as targets. The results confirmed the feasibility of the 2-D configuration and supported its integration into intelligent, real-time surveillance systems for security and underwater monitoring applications. - Open Access
Adversarial Robustness of Traffic Classification under Resource Constraints: Input Structure Matters
Traffic classification (TC) plays a critical role in cybersecurity, particularly in IoT and embedded contexts, where inspection must often occur locally under tight hardware constraints. We use hardware-aware neural architecture search (HW-NAS) to derive lightweight TC models that are accurate, efficient, and deployable on edge platforms. Two input formats are considered: a flattened byte sequence and a 2D packet-wise time series; we examine how input structure affects adversarial vulnerability when using resource-constrained models. Robustness is assessed against white-box attacks, specifically Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD). On USTC-TFC2016, both HW-NAS models achieve over 99% clean-data accuracy while remaining within 65k parameters and 2M FLOPs. Yet under perturbations of strength 0.1, their robustness diverges: the flat model retains over 85% accuracy, while the time-series variant drops below 35%. Adversarial fine-tuning delivers robust gains, with flat-input accuracy exceeding 96% and the time-series variant recovering over 60 percentage points in robustness, all without compromising efficiency. The results underscore how input structure influences adversarial vulnerability, and show that even compact, resource-efficient models can attain strong robustness, supporting their practical deployment in secure edge-based TC.
Open AccessAdversarial Robustness of Traffic Classification under Resource Constraints: Input Structure Matters
Traffic classification (TC) plays a critical role in cybersecurity, particularly in IoT and embedded contexts, where inspection must often occur locally under tight hardware constraints. We use hardware-aware neural architecture search (HW-NAS) to derive lightweight TC models that are accurate, efficient, and deployable on edge platforms. Two input formats are considered: a flattened byte sequence and a 2D packet-wise time series; we examine how input structure affects adversarial vulnerability when using resource-constrained models. Robustness is assessed against white-box attacks, specifically Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD). On USTC-TFC2016, both HW-NAS models achieve over 99% clean-data accuracy while remaining within 65k parameters and 2M FLOPs. Yet under perturbations of strength 0.1, their robustness diverges: the flat model retains over 85% accuracy, while the time-series variant drops below 35%. Adversarial fine-tuning delivers robust gains, with flat-input accuracy exceeding 96% and the time-series variant recovering over 60 percentage points in robustness, all without compromising efficiency. The results underscore how input structure influences adversarial vulnerability, and show that even compact, resource-efficient models can attain strong robustness, supporting their practical deployment in secure edge-based TC.
- Open Access
Energy-Efficient Deep Learning for Traffic Classification on Microcontrollers
In this paper, we present a practical deep learning (DL) approach for energy-efficient traffic classification (TC) on resource-limited microcontrollers, which are widely used in IoT-based smart systems and communication networks. Our objective is to balance accuracy, computational efficiency, and real-world deployability. To that end, we develop a lightweight 1D-CNN, optimized via hardware-aware neural architecture search (HW-NAS), which achieves 96.59% accuracy on the ISCX VPN-NonVPN dataset with only 88.26K parameters, a 20.12K maximum tensor size, and 10.08M floating-point operations (FLOPs). Moreover, it generalizes across various TC tasks, with accuracies ranging from 94% to 99%. To enable deployment, the model is quantized to INT8, suffering only a marginal 1-2% accuracy drop relative to its Float32 counterpart. We evaluate real-world inference performance on two microcontrollers: the high-performance STM32F746G-DISCO and the cost-sensitive Nucleo-F401RE. The deployed model achieves inference latencies of 31.43ms and 115.40ms, with energy consumption of 7.86 mJ and 29.10 mJ per inference, respectively. These results demonstrate the feasibility of on-device encrypted traffic analysis, paving the way for scalable, low-power IoT security solutions.
Open AccessEnergy-Efficient Deep Learning for Traffic Classification on Microcontrollers
In this paper, we present a practical deep learning (DL) approach for energy-efficient traffic classification (TC) on resource-limited microcontrollers, which are widely used in IoT-based smart systems and communication networks. Our objective is to balance accuracy, computational efficiency, and real-world deployability. To that end, we develop a lightweight 1D-CNN, optimized via hardware-aware neural architecture search (HW-NAS), which achieves 96.59% accuracy on the ISCX VPN-NonVPN dataset with only 88.26K parameters, a 20.12K maximum tensor size, and 10.08M floating-point operations (FLOPs). Moreover, it generalizes across various TC tasks, with accuracies ranging from 94% to 99%. To enable deployment, the model is quantized to INT8, suffering only a marginal 1-2% accuracy drop relative to its Float32 counterpart. We evaluate real-world inference performance on two microcontrollers: the high-performance STM32F746G-DISCO and the cost-sensitive Nucleo-F401RE. The deployed model achieves inference latencies of 31.43ms and 115.40ms, with energy consumption of 7.86 mJ and 29.10 mJ per inference, respectively. These results demonstrate the feasibility of on-device encrypted traffic analysis, paving the way for scalable, low-power IoT security solutions.
- 29 April 2023
- Published by Springer Nature
- p. 109-116
Affordance Segmentation Using RGB-D Sensors for Application in Portable Embedded Systems
- 29 April 2023
- Published by Springer Nature
- p. 109-116
Affordance Segmentation Using RGB-D Sensors for Application in Portable Embedded Systems
- Open Access
- 29 November 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (9) , 6753-6766
An approximate randomization-based neural network with dedicated digital architecture for energy-constrained devices
Variable energy constraints affect the implementations of neural networks on battery-operated embedded systems. This paper describes a learning algorithm for randomization-based neural networks with hard-limit activation functions. The approach adopts a novel cost function that balances accuracy and network complexity during training. From an energy-specific perspective, the new learning strategy allows to adjust, dynamically and in real time, the number of operations during the network’s forward phase. The proposed learning scheme leads to efficient predictors supported by digital architectures. The resulting digital architecture can switch to approximate computing at run time, in compliance with the available energy budget. Experiments on 10 real-world prediction testbeds confirmed the effectiveness of the learning scheme. Additional tests on limited-resource devices supported the implementation efficiency of the overall design approach.
- 29 November 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (9) , 6753-6766
Open AccessAn approximate randomization-based neural network with dedicated digital architecture for energy-constrained devices
Variable energy constraints affect the implementations of neural networks on battery-operated embedded systems. This paper describes a learning algorithm for randomization-based neural networks with hard-limit activation functions. The approach adopts a novel cost function that balances accuracy and network complexity during training. From an energy-specific perspective, the new learning strategy allows to adjust, dynamically and in real time, the number of operations during the network’s forward phase. The proposed learning scheme leads to efficient predictors supported by digital architectures. The resulting digital architecture can switch to approximate computing at run time, in compliance with the available energy budget. Experiments on 10 real-world prediction testbeds confirmed the effectiveness of the learning scheme. Additional tests on limited-resource devices supported the implementation efficiency of the overall design approach.
- Open Access
- 21 June 2022
- Published by Elsevier in Journal of Applied Geophysics
- Vol. 204
The LAMA system: A “smart” magnetometer network for harbour protection
- 21 June 2022
- Published by Elsevier in Journal of Applied Geophysics
- Vol. 204
Open AccessThe LAMA system: A “smart” magnetometer network for harbour protection
- Open Access
- 30 May 2022
- Published by MDPI AG in Biosensors
- Vol. 12 (6) , 375
From a Chemotherapeutic Drug to a High-Performance Nanocatalyst: A Fast Colorimetric Test for Cisplatin Detection at ppb Level
A rapid point-of-care method for the colorimetric detection of cisplatin was developed, exploiting the efficient conversion of the chemotherapeutic drug into a high-performance nanocatalyst with peroxidase enzyme mimics. This assay provides high specificity and ppb-detection sensitivity with the naked eye or a smartphone-based readout, outperforming many standard laboratory-based techniques. The nanocatalyst-enabled colorimetric assay can be integrated with machine-learning methods, providing accurate quantitative measurements. Such a combined approach opens interesting perspectives for the on-site monitoring of both chemotherapeutic patients to achieve optimal treatments and healthcare workers to prevent their unsafe exposure.
- 30 May 2022
- Published by MDPI AG in Biosensors
- Vol. 12 (6) , 375
Open AccessFrom a Chemotherapeutic Drug to a High-Performance Nanocatalyst: A Fast Colorimetric Test for Cisplatin Detection at ppb Level
A rapid point-of-care method for the colorimetric detection of cisplatin was developed, exploiting the efficient conversion of the chemotherapeutic drug into a high-performance nanocatalyst with peroxidase enzyme mimics. This assay provides high specificity and ppb-detection sensitivity with the naked eye or a smartphone-based readout, outperforming many standard laboratory-based techniques. The nanocatalyst-enabled colorimetric assay can be integrated with machine-learning methods, providing accurate quantitative measurements. Such a combined approach opens interesting perspectives for the on-site monitoring of both chemotherapeutic patients to achieve optimal treatments and healthcare workers to prevent their unsafe exposure.
- 28 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A pose-based hand image classification method retrainable on embedded target devices
State of the Art solutions for image classification are nowadays mostly based on Convolutional Neural Networks (CNNs), With respect to running them for inference, training CNNs is a much more resource demanding job. For this reason, embedded devices designed to execute neural networks are normally not suitable to train the same networks. Hand image classification is becoming increasingly popular in several application areas, by empowering embedded systems with the ability to recognise hand gestures or grasp types. In this field, making it possible to retrain the hand image classifier directly on the embedded device would allow an easy customization of classification task according to application specific needs. In this paper we present a method based on transfer learning which solves this problem by first extracting hand keypoints, using with a dedicated neural network, and then applying on the resulting data a lightweight classifier, easily retrainable directly on the target embedded system. The proposal has been validated on a new freely distributed dataset, outperforming state of the art solutions in the specific task.
- 28 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A pose-based hand image classification method retrainable on embedded target devices
State of the Art solutions for image classification are nowadays mostly based on Convolutional Neural Networks (CNNs), With respect to running them for inference, training CNNs is a much more resource demanding job. For this reason, embedded devices designed to execute neural networks are normally not suitable to train the same networks. Hand image classification is becoming increasingly popular in several application areas, by empowering embedded systems with the ability to recognise hand gestures or grasp types. In this field, making it possible to retrain the hand image classifier directly on the embedded device would allow an easy customization of classification task according to application specific needs. In this paper we present a method based on transfer learning which solves this problem by first extracting hand keypoints, using with a dedicated neural network, and then applying on the resulting data a lightweight classifier, easily retrainable directly on the target embedded system. The proposal has been validated on a new freely distributed dataset, outperforming state of the art solutions in the specific task.