Search Results (30)

Federated learning (FL) has emerged as a crucial technology in today’s data-centric environment, enabling decentralized machine learning while safeguarding user privacy. This study introduces “Federated Learning ML Operations (FedOps) Mobile”, a novel FL framework optimized for the dynamic and heterogeneous ecosystem of mobile devices. FedOps Mobile addresses the inherent challenges of FL—such as system scalability, device heterogeneity, and operational efficiency—through advanced on-device training using TensorFlow Lite and CoreML. The framework’s innovative approach includes sophisticated client selection mechanisms that assess device readiness and capabilities, ensuring equitable and efficient participation across the network. Additionally, FedOps Mobile leverages remote device control for seamless task management and continuous learning, all without compromising the user experience. The main contribution of this study is the demonstration that federated learning across heterogeneous devices, especially those using different operating systems, can be both practical and efficient using the FedOps Mobile framework. This was validated through experiments that evaluated three key areas: operational efficiency, model personalization, and resource optimization in multi-device settings. The results showed that the proposed method excels in client selection, energy consumption, and model optimization, establishing a new benchmark for federated learning in diverse and complex environments. Full article

In this paper, we introduce a security approach for on-device learning Edge AIs designed to detect abnormal conditions in factory machines. Since Edge AIs are easily accessible by an attacker physically, there are security risks due to physical attacks. In particular, there is a concern that the attacker may tamper with the training data of the on-device learning Edge AIs to degrade the task accuracy. Few risk assessments have been reported. It is important to understand these security risks before considering countermeasures. In this paper, we demonstrate a data poisoning attack against an on-device learning Edge AI. Our attack target is an on-device learning anomaly detection system. The system adopts MEMS accelerometers to measure the vibration of factory machines and detect anomalies. The anomaly detector also adopts a concept drift detection algorithm and multiple models to accommodate multiple normal patterns. For the attack, we used a method in which measurements are tampered with by exposing the MEMS accelerometer to acoustic waves of a specific frequency. The acceleration data falsified by this method were trained on an anomaly detector, and the result was that the abnormal state could not be detected. Full article

Data-driven models used for predictive classification and regression tasks are commonly computed using floating-point arithmetic and powerful computers. We address constraints in distributed sensor networks like the IoT, edge, and material-integrated computing, providing only low-resource embedded computers with sensor data that are acquired and processed locally. Sensor networks are characterized by strong heterogeneous systems. This work introduces and evaluates a virtual machine architecture that provides ML as a service layer (MLaaS) on the node level and addresses very low-resource distributed embedded computers (with less than 20 kB of RAM). The VM provides a unified ML instruction set architecture that can be programmed to implement decision trees, ANN, and CNN model architectures using scaled integer arithmetic only. Models are trained primarily offline using floating-point arithmetic, finally converted by an iterative scaling and transformation process, demonstrated in this work by two tests based on simulated and synthetic data. This paper is an extended version of the FedCSIS 2023 conference paper providing new algorithms and ML applications, including ANN/CNN-based regression and classification tasks studying the effects of discretization on classification and regression accuracy. Full article

Federated Learning (FL) is a decentralized machine learning method in which individual devices compute local models based on their data. In FL, devices periodically share newly trained updates with the central server, rather than submitting their raw data. The key characteristics of FL, including on-device training and aggregation, make it interesting for many communication domains. Moreover, the potential of new systems facilitating FL in sixth generation (6G) enabled Passive Optical Networks (PON), presents a promising opportunity for integration within this domain. This article focuses on the interaction between FL and PON, exploring approaches for effective bandwidth management, particularly in addressing the complexity introduced by FL traffic. In the PON standard, advanced bandwidth management is proposed by allocating multiple upstream grants utilizing the Dynamic Bandwidth Allocation (DBA) algorithm to be allocated for an Optical Network Unit (ONU). However, there is a lack of research on studying the utilization of multiple grant allocation. In this paper, we address this limitation by introducing a novel DBA approach that efficiently allocates PON bandwidth for FL traffic generation and demonstrates how multiple grants can benefit from the enhanced capacity of implementing PON in carrying out FL flows. Simulations conducted in this study show that the proposed solution outperforms state-of-the-art solutions in several network performance metrics, particularly in reducing upstream delay. This improvement holds great promise for enabling real-time data-intensive services that will be key components of 6G environments. Furthermore, our discussion outlines the potential for the integration of FL and PON as an operational reality capable of supporting 6G networking. Full article

This paper presents an on-device semi-supervised human activity detection system that can learn and predict human activity patterns in real time. The clinical objective is to monitor and detect the unhealthy sedentary lifestyle of a user. The proposed semi-supervised learning (SSL) framework uses sparsely labelled user activity events acquired from Inertial Measurement Unit sensors installed as wearable devices. The proposed cluster-based learning model in this approach is trained with data from the same target user, thus preserving data privacy while providing personalized activity detection services. Two different cluster labelling strategies, namely, population-based and distance-based strategies, are employed to achieve the desired classification performance. The proposed system is shown to be highly accurate and computationally efficient for different algorithmic parameters, which is relevant in the context of limited computing resources on typical wearable devices. Extensive experimentation and simulation study have been conducted on multi-user human activity data from the public domain in order to analyze the trade-off between classification accuracy and computation complexity of the proposed learning paradigm with different algorithmic hyper-parameters. With 4.17 h of training time for 8000 activity episodes, the proposed SSL approach consumes at most 20 KB of CPU memory space, while providing a maximum accuracy of 90% and 100% classification rates. Full article

In the context of TinyML, many research efforts have been devoted to designing forward topologies to support On-Device Learning. Reaching this target would bring numerous advantages, including reductions in latency and computational complexity, stronger privacy, data safety and robustness to adversarial attacks, higher resilience against concept drift, etc. However, On-Device Learning on resource constrained devices poses severe limitations to computational power and memory. Therefore, deploying Neural Networks on tiny devices appears to be prohibitive, since their backpropagation-based training is too memory demanding for their embedded assets. Using Extreme Learning Machines based on Convolutional Neural Networks might be feasible and very convenient, especially for Feature Extraction tasks. However, it requires searching for a randomly initialized topology that achieves results as good as those achieved by the backpropagated model. This work proposes a novel approach for automatically composing an Extreme Convolutional Feature Extractor, based on Neural Architecture Search and Bayesian Optimization. It was applied to the CIFAR-10 and MNIST datasets for evaluation. Two search spaces have been defined, as well as a search strategy that has been tested with two surrogate models, Gaussian Process and Random Forest. A performance estimation strategy was defined, keeping the feature set computed by the MLCommons-Tiny benchmark ResNet as a reference model. In as few as 1200 search iterations, the proposed strategy was able to achieve a topology whose extracted features scored a mean square error equal to 0.64 compared to the reference set. Further improvements are required, with a target of at least one order of magnitude decrease in mean square error for improved classification accuracy. The code is made available via GitHub to allow for the reproducibility of the results reported in this paper. Full article

The escalating global water usage and the increasing strain on major cities due to water shortages highlights the critical need for efficient water management practices. In water-stressed regions worldwide, significant water wastage is primarily attributed to leakages, inefficient use, and aging infrastructure. Undetected water leakages in buildings’ pipelines contribute to the water waste problem. To address this issue, an effective water leak detection method is required. In this paper, we explore the application of edge computing in smart buildings to enhance water management. By integrating sensors and embedded Machine Learning models, known as TinyML, smart water management systems can collect real-time data, analyze it, and make accurate decisions for efficient water utilization. The transition to TinyML enables faster and more cost-effective local decision-making, reducing the dependence on centralized entities. In this work, we propose a solution that can be adapted for effective leakage detection in real-world scenarios with minimum human intervention using TinyML. We follow an approach that is similar to a typical machine learning lifecycle in production, spanning stages including data collection, training, hyperparameter tuning, offline evaluation and model optimization for on-device resource efficiency before deployment. In this work, we considered an existing water leakage acoustic dataset for polyvinyl chloride pipelines. To prepare the acoustic data for analysis, we performed preprocessing to transform it into scalograms. We devised a water leak detection method by applying transfer learning to five distinct Convolutional Neural Network (CNN) variants, which are namely EfficientNet, ResNet, AlexNet, MobileNet V1, and MobileNet V2. The CNN models were found to be able to detect leakages where a maximum testing accuracy, recall, precision, and F1 score of 97.45%, 98.57%, 96.70%, and 97.63%, respectively, were observed using the EfficientNet model. To enable seamless deployment on the Arduino Nano 33 BLE edge device, the EfficientNet model is compressed using quantization resulting in a low inference time of 1932 ms, a peak RAM usage of 255.3 kilobytes, and a flash usage requirement of merely 48.7 kilobytes. Full article

Determining the cyclic-alternating-pattern (CAP) phases in sleep using electroencephalography (EEG) signals is crucial for assessing sleep quality. However, most current methods for CAP classification primarily rely on classical machine learning techniques, with limited implementation of deep-learning-based tools. Furthermore, these methods often require manual feature extraction. Herein, we propose a fully automatic deep-learning-based algorithm that leverages convolutional neural network architectures to classify the EEG signals via their time-frequency representations. Through our investigation, we explored using time-frequency analysis techniques and found that Wigner-based representations outperform the commonly used short-time Fourier transform for CAP classification. Additionally, our algorithm incorporates contextual information of the EEG signals and employs data augmentation techniques specifically designed to preserve the time-frequency structure. The model is developed using EEG signals of healthy subjects from the publicly available CAP sleep database (CAPSLPDB) on Physionet. An experimental study demonstrates that our algorithm surpasses existing machine-learning-based methods, achieving an accuracy of 77.5% on a balanced test set and 81.8% when evaluated on an unbalanced test set. Notably, the proposed algorithm exhibits efficiency and scalability, making it suitable for on-device implementation to enhance CAP identification procedures. Full article

Nowadays, the integration of home automation systems with smart thermostats is a common trend, designed to enhance resident comfort and conserve energy. The introduction of smart thermostats that can run machine learning algorithms has opened the door for on-device training, enabling customized thermal experiences in homes. However, leveraging the flexibility offered by on-device learning has been hindered by the absence of a tailored learning scheme that allows for accurate on-device training of thermal models. Traditional centralized learning (CL) and federated learning (FL) schemes rely on a central server that controls the learning experience, compromising the home’s privacy and requiring significant energy to operate. To address these challenges, we propose PePTM, a personalized peer-to-peer thermal modeling algorithm that generates tailored thermal models for each home, offering a controlled learning experience with a minimal training energy footprint while preserving the home’s privacy, an aspect difficult to achieve in both CL and FL. PePTM consists of local and collaborative learning phases that enable each home to train its thermal model and collaboratively improve it with a set of similar homes in a peer-to-peer fashion. To showcase the effectiveness of PePTM, we use a year’s worth of data from US homes to train thermal models using the RNN time-series model and compare the data across three learning schemes: CL, FL, and PePTM, in terms of model performance and the training energy footprint. Our experimental results show that PePTM is significantly energy-efficient, requiring 695 and 40 times less training energy than CL and FL, respectively, while maintaining comparable performance. We believe that PePTM sets the stage for new avenues for on-device thermal model training, providing a personalized thermal experience with reduced energy consumption and enhanced privacy. Full article

Hyperparameter optimization is one of the most tedious yet crucial steps in training machine learning models. There are numerous methods for this vital model-building stage, ranging from domain-specific manual tuning guidelines suggested by the oracles to the utilization of general purpose black-box optimization techniques. This paper proposes an agent-based collaborative technique for finding near-optimal values for any arbitrary set of hyperparameters (or decision variables) in a machine learning model (or a black-box function optimization problem). The developed method forms a hierarchical agent-based architecture for the distribution of the searching operations at different dimensions and employs a cooperative searching procedure based on an adaptive width-based random sampling technique to locate the optima. The behavior of the presented model, specifically against changes in its design parameters, is investigated in both machine learning and global function optimization applications, and its performance is compared with that of two randomized tuning strategies that are commonly used in practice. Moreover, we have compared the performance of the proposed approach against particle swarm optimization (PSO) and simulated annealing (SA) methods in function optimization to provide additional insights into its exploration in the search space. According to the empirical results, the proposed model outperformed the compared random-based methods in almost all tasks conducted, notably in a higher number of dimensions and in the presence of limited on-device computational resources. Full article

Recent advances both in hardware and software have facilitated the embedded intelligence (EI) research field, and enabled machine learning and decision-making integration in resource-scarce IoT devices and systems, realizing “conscious” and self-explanatory objects (smart objects). In the context of the broad use of WSNs in advanced IoT applications, this is the first work to provide an extreme-edge system, to address structural health monitoring (SHM) on polymethyl methacrylate (PPMA) thin-plate. To the best of our knowledge, state-of-the-art solutions primarily utilize impact positioning methods based on the time of arrival of the stress wave, while in the last decade machine learning data analysis has been performed, by more expensive and resource-abundant equipment than general/development purpose IoT devices, both for the collection and the inference stages of the monitoring system. In contrast to the existing systems, we propose a methodology and a system, implemented by a low-cost device, with the benefit of performing an online and on-device impact localization service from an agnostic perspective, regarding the material and the sensors’ location (as none of those attributes are used). Thus, a design of experiments and the corresponding methodology to build an experimental time-series dataset for impact detection and localization is proposed, using ceramic piezoelectric transducers (PZTs). The system is excited with a steel ball, varying the height from which it is released. Based on TinyML technology for embedding intelligence in low-power devices, we implement and validate random forest and shallow neural network models to localize in real-time (less than 400 ms latency) any occurring impacts on the structure, achieving higher than 90% accuracy. Full article

Embedded machine learning, TinyML, is a relatively new and fast-growing field of ML, enabling on-device sensor data analytics at low power requirements. This paper presents possible improvements to GMOS, a gas sensor, using TinyML technology. GMOS is a low-cost catalytic gas sensor, fabricated with the standard CMOS-SOI process, based on a suspended thermal transistor MOS (TMOS). Exothermic combustion reactions lead to temperature increases, which modify the suspended transistor’s (used as the sensing element) current-voltage characteristics. We were able to use GMOS measurements for gas classification (both for gas types, as well as concentration), resulting in high-proficiency gas detection at a low cost. Our preliminary results show great successes in the detection of ethanol and acetone gases. Moreover, we believe the method could be generalized to more gas types, concentrations, and gas mixes in future research. Full article

Personalised healthcare has seen significant improvements due to the introduction of health monitoring technologies that allow wearable devices to unintrusively monitor physiological parameters such as heart health, blood pressure, sleep patterns, and blood glucose levels, among others. Additionally, utilising advanced sensing technologies based on flexible and innovative biocompatible materials in wearable devices allows high accuracy and precision measurement of biological signals. Furthermore, applying real-time Machine Learning algorithms to highly accurate physiological parameters allows precise identification of unusual patterns in the data to provide health event predictions and warnings for timely intervention. However, in the predominantly adopted architectures, health event predictions based on Machine Learning are typically obtained by leveraging Cloud infrastructures characterised by shortcomings such as delayed response times and privacy issues. Fortunately, recent works highlight that a new paradigm based on Edge Computing technologies and on-device Artificial Intelligence significantly improve the latency and privacy issues. Applying this new paradigm to personalised healthcare architectures can significantly improve their efficiency and efficacy. Therefore, this paper reviews existing IoT healthcare architectures that utilise wearable devices and subsequently presents a scalable and modular system architecture to leverage emerging technologies to solve identified shortcomings. The defined architecture includes ultrathin, skin-compatible, flexible, high precision piezoelectric sensors, low-cost communication technologies, on-device intelligence, Edge Intelligence, and Edge Computing technologies. To provide development guidelines and define a consistent reference architecture for improved scalable wearable IoT-based critical healthcare architectures, this manuscript outlines the essential functional and non-functional requirements based on deductions from existing architectures and emerging technology trends. The presented system architecture can be applied to many scenarios, including ambient assisted living, where continuous surveillance and issuance of timely warnings can afford independence to the elderly and chronically ill. We conclude that the distribution and modularity of architecture layers, local AI-based elaboration, and data packaging consistency are the more essential functional requirements for critical healthcare application use cases. We also identify fast response time, utility, comfort, and low cost as the essential non-functional requirements for the defined system architecture. Full article

Currently, Android is the most popular operating system among mobile devices. However, as the number of devices with the Android operating system increases, so does the danger of using them. This is especially important as smartphones increasingly authenticate critical activities(e-banking, e-identity). BotSense Mobile is a tool already integrated with some critical applications (e-banking, e-identity) to increase user safety. In this paper, we focus on the novel functionality of BotSense Mobile: the detection of malware applications on a user device. In addition to the standard blacklist approach, we propose a machine learning-based model for unknown malicious application detection. The lightweight neural network model is deployed on an edge device to avoid sending sensitive user data outside the device. For the same reason, manifest-related features can be used by the detector only. We present a comprehensive empirical analysis of malware detection conducted on recent data (May–June, 2022) from the Koodous platform, which is a collaborative platform where over 70 million Android applications were collected. The research highlighted the problem of machine learning model aging. We evaluated the lightweight model on recent Koodous data and obtained $f_1=0.77$ and high precision ($0.9$). Full article

In smart grids (SGs), the systematic utilization of consumer energy data while maintaining its privacy is of paramount importance. This research addresses this problem by energy theft detection while preserving the privacy of client data. In particular, this research identifies centralized models as more accurate in predicting energy theft in SGs but with no or significantly less data protection. Current research proposes a novel federated learning (FL) framework, namely FedDP, to tackle this issue. The proposed framework enables various clients to benefit from on-device prediction with very little communication overhead and to learn from the experience of other clients with the help of a central server (CS). Furthermore, for the accurate identification of energy theft, the use of a novel federated voting classifier (FVC) is proposed. FVC uses the majority voting-based consensus of traditional machine learning (ML) classifiers namely, random forests (RF), k-nearest neighbors (KNN), and bagging classifiers (BG). To the best of our knowledge, conventional ML classifiers have never been used in a federated manner for energy theft detection in SGs. Finally, substantial experiments are performed on the real-world energy consumption dataset. Results illustrate that the proposed model can accurately and efficiently detect energy theft in SGs while guaranteeing the security of client data. Full article