Adaptive Federated Learning in Resource Constrained Edge Computing Systems

Emerging technologies and applications including Internet of Things (IoT), social networking, and crowd-sourcing generate large amounts of data at the network edge. Machine learning models are often built from the collected data, to enable the detection, classification, and prediction of future events. Due to bandwidth, storage, and privacy concerns, it is often impractical to send all the data to a centralized location. In this paper, we consider the problem of learning model parameters from data distributed across multiple edge nodes, without sending raw data to a centralized place. Our focus is on a generic class of machine learning models that are trained using gradient-descent based approaches. We analyze the convergence rate of distributed gradient descent from a theoretical point of view, based on which we propose a control algorithm that determines the best trade-off between local update and global parameter aggregation to minimize the loss function under a given resource budget. The performance of the proposed algorithm is evaluated via extensive experiments with real datasets, both on a networked prototype system and in a larger-scale simulated environment. The experimentation results show that our proposed approach performs near to the optimum with various machine learning models and different data distributions.

We study the performance of classical and quantum machine learning (ML) models in predicting outcomes of physical experiments. The experiments depend on an input parameter x and involve execution of a (possibly unknown) quantum process E. Our figure of merit is the number of runs of E required to achieve a desired prediction performance. We consider classical ML models that perform a measurement and record the classical outcome after each run of E, and quantum ML models that can access E coherently to acquire quantum data; the classical or quantum data are then used to predict the outcomes of future experiments. We prove that for any input distribution D(x), a classical ML model can provide accurate predictions on average by accessing E a number of times comparable to the optimal quantum ML model. In contrast, for achieving an accurate prediction on all inputs, we prove that the exponential quantum advantage is possible. For example, to predict the expectations of all Pauli observables in an n-qubit system ρ, classical ML models require 2^{Ω(n)} copies of ρ, but we present a quantum ML model using only O(n) copies. Our results clarify where the quantum advantage is possible and highlight the potential for classical ML models to address challenging quantum problems in physics and chemistry.

SummarySpatiotemporal solar radiation forecasting is extremely challenging due to its dependence on metrological and environmental factors. Chaotic time‐varying and non‐linearity make the forecasting model more complex. To cater this crucial issue, the paper provides a comprehensive investigation of the deep learning framework for the prediction of the two components of solar irradiation, that is, Diffuse Horizontal Irradiance (DHI) and Direct Normal Irradiance (DNI). Through exploratory data analysis the three recent most prominent deep learning (DL) architecture have been developed and compared with the other classical machine learning (ML) models in terms of the statistical performance accuracy. In our study, DL architecture includes convolutional neural network (CNN) and recurrent neural network (RNN) whereas classical ML models include Random Forest (RF), Support Vector Regression (SVR), Multilayer Perceptron (MLP), Extreme Gradient Boosting (XGB), and K‐Nearest Neighbor (KNN). Additionally, three optimization techniques Grid Search (GS), Random Search (RS), and Bayesian Optimization (BO) have been incorporated for tuning the hyper parameters of the classical ML models to obtain the best results. Based on the rigorous comparative analysis it was found that the CNN model has outperformed all classical machine learning and DL models having lowest mean squared error and highest R‐Squared value with least computational time.

Bronchiolitis remains a frequent reason for hospitalization in infants during the winter season. Epidemiologic surveillance remains crucial in the era of widespread immunoprophylaxis for the leading viral agent causing bronchiolitis. We investigated the performance of classical machine learning (ML) models, Deep Learning (DL), and a pre-trained large language model (LLM) in classifying bronchiolitis diagnosis from the free-text-diagnosis field of the emergency department electronic health records (EHRs). As a secondary aim, we evaluated the diagnostic accuracy of the actual official administrative ICD-9 encoding for Bronchiolitis diagnosis. 28,557 records of infants < 1 year with complete discharge diagnoses fields were retrieved between the years 2007–2018 and manually classified by an expert pediatrician to create the gold standard diagnosis set for training the algorithm. After data pre-processing, classical ML models (Random Forest, Decision Tree, Gradient Boosting Machine, Linear Discriminant Analysis, Support Vector Machine), a Deep Learning (DL) tool, and a pre-trained LLM (GPT-5) were evaluated using balanced accuracy, sensitivity, and F1 scores. The official administrative ICD-9 encoding classification accuracy was compared to the gold standard. Overall, 1,903 of 28,557 records (6.7%) were classified as bronchiolitis by the gold standard approach. The DL model and GPT-5 outperformed traditional ML models, achieving higher sensitivities (0.97, 95%CI 0.96-1.00, and 0.98, 95% CI 0.98–0.99, respectively), F1 scores (0.96, 95% CI 0.95–0.99, and 0.99, 95% CI 0.98–0.99, respectively), and balanced accuracy (0.98, 95%CI 0.98-1.00, and 0.99, 95% CI 0.99–0.99, respectively). Traditional ML models showed sensitivities between 0.77 and 0.98, F1 scores between 0.86 and 0.96, and balanced accuracies between 0.88 and 0.96. ICD-9 codes showed sensitivity of 85.9% (95% CI 84.27–87.45), and specificity of 98.5% (95% CI 98.36–98.65). To our knowledge, this is the first study directly comparing an LLM, deep learning, and multiple classical ML models for bronchiolitis surveillance in the post-Nirsevimab era. DL and GPT-5 outperformed traditional ML-based tools in identifying bronchiolitis diagnoses and ICD-9 diagnosis coding. AI-based tools hold significant potential for improving epidemiologic surveillance of bronchiolitis from emergency department EHRs. Not applicable.

Improving autocoding performance of rare categories in injury classification: Is more training data or filtering the solution?

Stream temperature (Ts) is an important water quality parameter that affects ecosystem health and human water use for beneficial purposes. Accurate Ts predictions at different spatial and temporal scales can inform water management decisions that account for the effects of changing climate and extreme events. In particular, widespread predictions of Ts in unmonitored stream reaches can enable decision makers to be responsive to changes caused by unforeseen disturbances. In this study, we demonstrate the use of classical machine learning (ML) models, support vector regression and gradient boosted trees (XGBoost), for monthly Ts predictions in 78 pristine and human-impacted catchments of the Mid-Atlantic and Pacific Northwest hydrologic regions spanning different geologies, climate, and land use. The ML models were trained using long-term monitoring data from 1980–2020 for three scenarios: (1) temporal predictions at a single site, (2) temporal predictions for multiple sites within a region, and (3) spatiotemporal predictions in unmonitored basins (PUB). In the first two scenarios, the ML models predicted Ts with median root mean squared errors (RMSE) of 0.69–0.84 °C and 0.92–1.02 °C across different model types for the temporal predictions at single and multiple sites respectively. For the PUB scenario, we used a bootstrap aggregation approach using models trained with different subsets of data, for which an ensemble XGBoost implementation outperformed all other modeling configurations (median RMSE 0.62 °C).The ML models improved median monthly Ts estimates compared to baseline statistical multi-linear regression models by 15–48% depending on the site and scenario. Air temperature was found to be the primary driver of monthly Ts for all sites, with secondary influence of month of the year (seasonality) and solar radiation, while discharge was a significant predictor at only 10 sites. The predictive performance of the ML models was robust to configuration changes in model setup and inputs, but was influenced by the distance to the nearest dam with RMSE &lt;1 °C at sites situated greater than 16 and 44 km from a dam for the temporal single site and regional scenarios, and over 1.4 km from a dam for the PUB scenario. Our results show that classical ML models with solely meteorological inputs can be used for spatial and temporal predictions of monthly Ts in pristine and managed basins with reasonable (&lt;1 °C) accuracy for most locations.

The proliferation of invasive plant species poses a significant ecological threat, necessitating effective mapping strategies for control and conservation efforts. Existing studies employing unmanned aerial vehicles (UAVs) and multispectral (MS) sensors in complex natural environments have predominantly relied on classical machine learning (ML) models for mapping plant species in natural environments. However, a critical gap exists in the literature regarding the use of deep learning (DL) techniques that integrate MS data and vegetation indices (VIs) with different feature extraction techniques to map invasive species in complex natural environments. This research addresses this gap by focusing on mapping the distribution of the Broad-leaved pepper (BLP) along the coastal strip in the Sunshine Coast region of Southern Queensland in Australia. The methodology employs a dual approach, utilising classical ML models including Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Support Vector Machine (SVM) in conjunction with the U-Net DL model. This comparative analysis allows for an in-depth evaluation of the performance and effectiveness of both classical ML and advanced DL techniques in mapping the distribution of BLP along the coastal strip. Results indicate that the DL U-Net model outperforms classical ML models, achieving a precision of 83%, recall of 81%, and F1–score of 82% for BLP classification during training and validation. The DL U-Net model attains a precision of 86%, recall of 76%, and F1–score of 81% for BLP classification, along with an Intersection over Union (IoU) of 68% on the separate test dataset not used for training. These findings contribute valuable insights to environmental conservation efforts, emphasising the significance of integrating MS data with DL techniques for the accurate mapping of invasive plant species.

This paper aims to compare the performance of the classical machine learning (CML) model and the deep learning (DL) model, and to assess the effectiveness of utilizing fusion radiomics from both CML and DL in distinguishing encephalitis from glioma in atypical cases. We analysed the axial FLAIR images of preoperative MRI in 116 patients pathologically confirmed as gliomas and clinically diagnosed with encephalitis. The 3 CML models (logistic regression (LR), support vector machine (SVM) and multi-layer perceptron (MLP)), 3 DL models (DenseNet 121, ResNet 50 and ResNet 18) and a deep learning radiomic (DLR) model were established, respectively. The area under the receiver operating curve (AUC) and sensitivity, specificity, accuracy, negative predictive value (NPV) and positive predictive value (PPV) were calculated for the training and validation sets. In addition, a deep learning radiomic nomogram (DLRN) and a web calculator were designed as a tool to aid clinical decision-making. The best DL model (ResNet50) consistently outperformed the best CML model (LR). The DLR model had the best predictive performance, with AUC, sensitivity, specificity, accuracy, NPV and PPV of 0.879, 0.929, 0.800, 0.875, 0.867 and 0.889 in the validation sets, respectively. Calibration curve of DLR model shows good agreement between prediction and observation, and the decision curve analysis (DCA) indicated that the DLR model had higher overall net benefit than the other two models (ResNet50 and LR). Meanwhile, the DLRN and web calculator can provide dynamic assessments. Machine learning (ML) models have the potential to non-invasively differentiate between encephalitis and glioma in atypical cases. Furthermore, combining DL and CML techniques could enhance the performance of the ML models.

Federated Learning (FL) is a distributed learning methodology that allows multiple nodes to cooperatively train a deep learning model, without the need to share their local data. It is a promising solution for telemonitoring systems that demand intensive data collection, for detection, classification, and prediction of future events, from different locations while maintaining a strict privacy constraint. Due to privacy concerns and critical communication bottlenecks, it can become impractical to send the FL updated models to a centralized server. Thus, this paper studies the potential of hierarchical FL in Internet of Things (IoT) heterogeneous systems. In particular, we propose an optimized solution for user assignment and resource allocation over hierarchical FL architecture for IoT heterogeneous systems. This work focuses on a generic class of machine learning models that are trained using gradient-descent-based schemes while considering the practical constraints of non-uniformly distributed data across different users. We evaluate the proposed system using two real-world datasets, and we show that it outperforms state-of-the-art FL solutions. Specifically, our numerical results highlight the effectiveness of our approach and its ability to provide 4–6% increase in the classification accuracy, with respect to hierarchical FL schemes that consider distance-based user assignment. Furthermore, the proposed approach could significantly accelerate FL training and reduce communication overhead by providing 75–85% reduction in the communication rounds between edge nodes and the centralized server, for the same model accuracy.

Lung adenocarcinoma (LUAD) exhibits molecular heterogeneity, with mitochondrial damage affecting progression. The relationship between mitochondrial damage and immune infiltration, and Weighted Gene Co-expression Network Analysis (WGCNA)-derived biomarkers for LUAD classification and prognosis, remains unexplored. The objective of our research is to identify gene modules closely related to the clinical stages of LUAD using the WGCNA method. Based on the genes within these modules, we constructed machine learning (ML) models for classification and prognosis prediction, thereby facilitating precise diagnosis and personalized treatment of LUAD. Using GeneCards and The Cancer Genome Atlas (TCGA) databases, we screened differentially expressed mitochondrial damage-related genes in LUAD. Immune cell infiltration patterns were assessed using Single-Sample Gene Set Enrichment Analysis (SSGSEA) method. Functional enrichment analyses were conducted to explore biological functions and signaling pathways. Gene modules related to clinical stages of LUAD were identified by WGCNA. ML models were constructed for classification and prognosis prediction, and validated in an independent Gene Expression Omnibus (GEO) dataset. The study revealed a significant relationship between mitochondrial damage and immune infiltration in LUAD. We identified a gene module closely associated with the clinical stages of LUAD. The ML models for classification and prognosis that were constructed demonstrated good effectiveness and generalization capabilities. Mitochondrial damage-related genes are crucial in LUAD progression and linked to immune infiltration. The gene module and models identified have potential applications in LUAD classification and prognosis, offering novel markers for precision medicine. This study uncovers the relationship between mitochondrial damage and immune infiltration in LUAD, paving the way for molecular classification, prognosis prediction, and personalized treatment strategies.

Heart failure is a devastating disease that has high mortality rates and a negative impact on quality of life. Heart failure patients often experience emergency readmission after an initial episode, often due to inadequate management. A timely diagnosis and treatment of underlying issues can significantly reduce the risk of emergency readmissions. The purpose of this project was to predict emergency readmissions of discharged heart failure patients using classical machine learning (ML) models based on Electronic Health Record (EHR) data. The dataset used for this study consisted of 166 clinical biomarkers from 2008 patient records. Three feature selection techniques were studied along with 13 classical ML models using five-fold cross-validation. A stacking ML model was trained using the predictions of the three best-performing models for final classification. The stacking ML model provided an accuracy, precision, recall, specificity, F1-score, and area under the curve (AUC) of 89.41%, 90.10%, 89.41%, 87.83%, 89.28%, and 0.881, respectively. This indicates the effectiveness of the proposed model in predicting emergency readmissions. The healthcare providers can intervene pro-actively to reduce emergency hospital readmission risk and improve patient outcomes and decrease healthcare costs using the proposed model.

Cooperative Swarm Learning for Distributed Cyclic Edge Intelligent Computing

PurposeTo develop machine learning (ML) models to predict, at baseline, treatment outcomes at month 9 in patients with neovascular age-related macular degeneration (nAMD) receiving faricimab.DesignRetrospective proof of concept study.ParticipantsPatients enrolled in the phase II AVENUE trial (NCT02484690) of faricimab in nAMD.MethodsBaseline characteristics and spectral domain-OCT (SD-OCT) image data from 185 faricimab-treated eyes were split into 80% training and 20% test sets at the patient level. Input variables were baseline age, sex, best-corrected visual acuity (BCVA), central subfield thickness (CST), low luminance deficit, treatment arm, and SD-OCT images. A regression problem (BCVA) and a binary classification problem (reduction of CST by 35%) were considered. Overall, 10 models were developed and tested for each problem. Benchmark classical ML models (linear, random forest, extreme gradient boosting) were trained on baseline characteristics; benchmark deep neural networks (DNNs) were trained on baseline SD-OCT B-scans. Baseline characteristics and SD-OCT data were merged using 2 approaches: model stacking (using DNN prediction as an input feature for classical ML models) and model averaging (which averaged predictions from the DNN using SD-OCT volume and from classical ML models using baseline characteristics).Main Outcome MeasuresTreatment outcomes were defined by 2 target variables: functional (BCVA letter score) and anatomical (percent decrease in CST from baseline) outcomes at month 9.ResultsThe best-performing BCVA regression model with respect to the test coefficient of determination (R2) was the linear model in the model-stacking approach with R2 of 0.31. The best-performing CST classification model with respect to test area under receiver operating characteristics (AUROC) was the benchmark linear model with AUROC of 0.87. A post hoc analysis showed the baseline BCVA and the baseline CST had the most effect in the all-model prediction for BCVA regression and CST classification, respectively.ConclusionsPromising signals for predicting treatment outcomes from baseline characteristics were detected; however, the predictive benefit of baseline images was unclear in this proof-of-concept study. Further testing and validation with larger, independent datasets is required to fully explore the predictive capacity of ML models using baseline imaging data.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

COVID-19 is highly infectious and causes acute respiratory disease. Machine learning (ML) and deep learning (DL) models are vital in detecting disease from computerized chest tomography (CT) scans. The DL models outperformed the ML models. For COVID-19 detection from CT scan images, DL models are used as end-to-end models. Thus, the performance of the model is evaluated for the quality of the extracted feature and classification accuracy. There are four contributions included in this work. First, this research is motivated by studying the quality of the extracted feature from the DL by feeding these extracted to an ML model. In other words, we proposed comparing the end-to-end DL model performance against the approach of using DL for feature extraction and ML for the classification of COVID-19 CT scan images. Second, we proposed studying the effect of fusing extracted features from image descriptors, e.g., Scale-Invariant Feature Transform (SIFT), with extracted features from DL models. Third, we proposed a new Convolutional Neural Network (CNN) to be trained from scratch and then compared to the deep transfer learning on the same classification problem. Finally, we studied the performance gap between classic ML models against ensemble learning models. The proposed framework is evaluated using a CT dataset, where the obtained results are evaluated using five different metrics The obtained results revealed that using the proposed CNN model is better than using the well-known DL model for the purpose of feature extraction. Moreover, using a DL model for feature extraction and an ML model for the classification task achieved better results in comparison to using an end-to-end DL model for detecting COVID-19 CT scan images. Of note, the accuracy rate of the former method improved by using ensemble learning models instead of the classic ML models. The proposed method achieved the best accuracy rate of 99.39%.

BackgroundThe aim of this study was to develop a machine learning (ML) model for classifying osteoporosis in Korean women based on a large-scale population cohort study. This study also aimed to assess ML model performance compared with traditional osteoporosis screening tools. Furthermore, this study aimed to examine the factors influencing the risk of osteoporosis through variable importance.MethodsData was collected from 4199 women aged 40–69 years in the baseline survey of the Ansan and Ansung cohort of the Korean Genome and Epidemiology Study. Osteoporosis was set as the dependent variable to develop ML classification models. Independent variables included 122 factors related to osteoporosis risk, such as socio-demographic characteristics, anthropometric parameters, lifestyle factors, reproductive factors, nutrient intakes, diet quality indices, medical history, medication history, family history, biochemical parameters, and genetic factors. The six classification models were developed using ML techniques, including decision tree, random forest, multilayer perceptron, support vector machine, light gradient boosting machine, and extreme gradient boosting (XGBoost). The six ML classification models were compared with two traditional osteoporosis screening tools, including the osteoporosis risk assessment instrument (ORAI) and the osteoporosis self-assessment tool (OST). The ML model performances were evaluated and compared using the confusion matrix and area under the curve (AUC) metrics. Variable importance was assessed using the XGBoost technique to investigate osteoporosis risk factors.ResultsThe XGBoost model showed the highest performance out of the six ML classification models, with an accuracy of 0.705, precision of 0.664, recall of 0.830, and F1 score of 0.738. Moreover, the XGBoost model showed a higher performance on AUC than ORAI and OST. Variable importance scores were identified for 69 out of the 122 variables associated with osteoporosis risk factors. Age at menopause ranked first in variable importance. Variables of arthritis, physical activities, hypertension, education level, income level; alcohol intake, potassium intake, homeostatic model assessment for insulin resistance; energy intake, vitamin C intake, gout; and dietary inflammatory index ranked in the top 20 out of the 69 variables, using the XGBoost technique.ConclusionsThis study found that an XGBoost model can be utilized to classify osteoporosis in Korean women. Age at menopause is a significant factor in osteoporosis risk, followed by arthritis, physical activities, hypertension, and education level.