Aggregation Method Research Articles

Industrial data-driven machine learning soft sensing for optimal operation of etching tools

Smart Manufacturing, or Industry 4.0, has gained significant attention in recent decades with the integration of Internet of Things (IoT) and Information Technologies (IT). As modern production methods continue to increase in complexity, there is a greater need to consider what variables can be physically measured. This advancement necessitates the use of physical sensors to comprehensively and directly gather measurable data on industrial processes; specifically, these sensors gather data that can be recontextualized into new process information. For example, artificial intelligence (AI) machine learning-based soft sensors can increase operational productivity and machine tool performance while still ensuring that critical product specifications are met. One industry that has a high volume of labor-intensive, time-consuming, and expensive processes is the semiconductor industry. AI machine learning methods can meet these challenges by taking in operational data and extracting process-specific information needed to meet the high product specifications of the industry. However, a key challenge is the availability of high quality data that covers the full operating range, including the day-to-day variance. This paper examines the applicability of soft sensing methods to the operational data of five industrial etching machines. Data is collected from readily accessible and cost-effective physical sensors installed on the tools that manage and control the operating conditions of the tool. The operational data are then used in an intelligent data aggregation approach that increases the scope and robustness for soft sensors in general by creating larger training datasets comprised of high value data with greater operational ranges and process variation. The generalized soft sensor can then be fine-tuned and validated for a particular machine. In this paper, we test the effects of data aggregation for high performing Feedforward Neural Network (FNN) models that are constructed in two ways: first as a classifier to estimate product PASS/FAIL outcomes and second as a regressor to quantitatively estimate oxide thickness. For PASS/FAIL classification, a data aggregation method is developed to enhance model predictive performance with larger training datasets. A statistical analysis method involving point-biserial correlation and the Mean Absolute Error (MAE) difference score is introduced to select the optimal candidate datasets for aggregation, further improving the effectiveness of data aggregation. For large datasets with high quality data that enable model training for more complex tasks, regression models that predict the oxide thickness of the product are also developed. Two types of models with different loss functions are tested to compare the effects of the Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE) loss functions on model performance. Both the classification and regression models can be applied in industrial settings as they provide additional information regarding the process outcome. Individually, these models can reduce the number of metrology steps in semiconductor factories, and when developed further, can empower the development of advanced process control strategies.

Language-based reasoning graph neural network for commonsense question answering

Language model (LM) has played an increasingly important role in the common-sense understanding and reasoning in the CSQA task (Common Sense Question Answering). However, due to the amount of model parameters, increasing training data helps little in further improving model performance. Introducing external knowledge through graph neural networks (GNNs) proves positive in boosting performance, but exploiting different knowledge sources and capturing contextual information between text and knowledge inside remains a challenge. In this paper, we propose LBR-GNN, a Language-Based Reasoning Graph Neural Network method to address these problems, by representing the question with each answer and external knowledge using a language model and predicting the reasoning score with a designed language-based GNN. Our LBR-GNN will first regulate external knowledge into a consistent textual form and encode it using a standard LM to capture the contextual information. Then, we build a graph neural network using the encoded information, especially the language-level edge representation. Finally, we design a novel edge aggregation method to select the edge information for GNN update and the language-guided GNN reasoning. We assess the performance of LBR-GNN across the CommonsenseQA, CommonsenseQA-IH, and OpenBookQA datasets. Our evaluation reveals a performance boost of more than 5% compared to the state-of-the-art methods on the CSQA dataset, achieved with a similar number of additional parameters.

DEVELOPMENT OF A METHODOLOGY FOR DATA NORMALISATION AND AGGREGATION TO ENHANCE SECURITY LEVELS IN INTERNET OF THINGS INTERACTIONS

The number of interacting devices is increasing every day, and with this constant innovation, serious security challenges arise. The concept of the Internet of Things is being actively applied in both domestic and industrial settings. Researchers are increasingly highlighting the challenges and importance of network security. Data preprocessing plays an important role in security by transforming the input data corresponding to algorithmic criteria and thereby contributing to the prediction accuracy. The data preprocessing process is determined by many factors, including the processing algorithm, the data, and the application. Moreover, in Internet of Things interactions, data normalisation and aggregation can significantly improve security and reduce the amount of data used further decision making. This paper discusses the challenges of data normalisation and aggregation in the IoT to handle large amounts of data generated by multiple connected IoT devices. A secure data normalisation and aggregation method promotes successful minimised data transfer over the network and provides scalability to meet the increasing demands of IoT deployment. The proposed work presents approaches used in data aggregation protocols that address interference, fault tolerance, security and mobility issues. A local aggregation approach using the run-length encoding algorithm is presented. The proposed technique consists of data acquisition, data preprocessing, data normalisation and data aggregation steps. Data normalisation was performed via the Z-score algorithm, and the LEACH algorithm was used for data aggregation. In the experimental study, the percentage of faulty nodes reached 35%. The performance of the proposed solution was 0.82. The results demonstrate a reduction in resource consumption while maintaining the value and integrity of the data.

A Secure and Fair Federated Learning Framework Based on Consensus Incentive Mechanism

Federated learning facilitates collaborative computation among multiple participants while safeguarding user privacy. However, current federated learning algorithms operate under the assumption that all participants are trustworthy and their systems are secure. Nonetheless, real-world scenarios present several challenges: (1) Malicious clients disrupt federated learning through model poisoning and data poisoning attacks. Although some research has proposed secure aggregation methods to address this issue, many methods have inherent limitations. (2) Clients may refuse or passively participate in the training process due to considerations of self-interest, and may even interfere with the training process due to competitive relationships. To overcome these obstacles, we have devised a reliable federated framework aimed at ensuring secure computing throughout the entirety of federated task processes. Initially, we propose a method for detecting malicious models to safeguard the integrity of model aggregation. Furthermore, we have proposed a fair contribution assessment method and awarded the right to write blocks to the creator of the optimal model, ensuring the active participation of participants in both local training and model aggregation. Finally, we establish a computational framework grounded in blockchain and smart contracts to uphold the integrity and fairness of federated tasks. To assess the efficacy of our framework, we conduct simulations involving various types of client attacks and contribution assessment scenarios using multiple open-source datasets. Results from these experiments demonstrate that our framework effectively ensures the credibility of federated tasks while achieving impartial evaluation of client contributions.

Assessment of the Climate Change Vulnerability of the Cities in Turkey

Climate-related vulnerability indices are increasingly being utilized to enhance the creation of better disaster management strategies and to better understand and anticipate the effects of disasters related to climate change. This study evaluates the climate change vulnerability of the cities in Türkiye through focusing on their exposure, susceptibilities and adaptive capacities to climate change. Data from social, economic and environmental sub-indicators were assessed and most relevant indicators were aggregated with the goal of constructing a composite Climate Change Vulnerability Index (CCVI). The CCVI includes six forms of capital leading to socio-economic and environmental sustainability i.e. social, public utility and transport, economics, land cover, meteorology and atmospheric conditions, and natural disaster capitals, and will be assessed combining each of these forms of capitals and its three dimensions (exposure, susceptibility and adaptive capacity). Stakeholder-driven structured methodology that discovers and ranks context-relevant indicators and sets weights for aggregating indicator scores by using the Best-Worst method (BWM) and stepwise weight assessment ratio analysis (SWARA) method are utilised. The indicators are aggregated through application of the BWM and SWARA weights using a linear aggregation method. From BWM and SWARA, the highest weights were obtained for meteorological conditions and land cover which are more than 0.36 and 0.22, respectively. The lowest weights were assigned to social characteristics and economy both of which were smaller than 0.10. The findings indicated that the cities on the northern, western and southern coasts as well as the cities in south-east region are the most vulnerable to climate change. The construction of CCVI can be used as part of decision-making process to minimize hazards and exposure to risk of climate change for the cities of Türkiye.

Aggregative Swab Sampling Method for Romaine Lettuce Show Similar Quality and Safety Indicators and Microbial Profiles Compared to Composite Produce Leaf Samples in a Pilot Study.

Composite produce leaf samples from commercial production rarely test positive for pathogens, potentially due to low pathogen prevalence or the relatively small number of plants sampled. Aggregative sampling may offer a more representative alternative. This pilot study investigated whether aggregative swab samples performed similarly to produce leaf samples in their ability to recover quality indicators (APCs and coliforms), detect Escherichia coli, and recover representative microbial profiles. Aggregative swabs of the outer leaves of romaine plants (n = 12) and composite samples consisting of various grabs of produce leaves (n = 14) were collected from 60 by 28 ft sections of a one-acre commercial romaine lettuce field. Aerobic plate counts were 9.17 ± 0.43 and 9.21 ± 0.42 log(CFU/g) for produce leaf samples and swabs, respectively. Means and variance were not significantly different (p = 0.38 and p = 0.92, respectively). Coliform recoveries were 3.80 ± 0.76 and 4.19 ± 1.15 log(CFU/g) for produce leaf and swabs, respectively. Means and variances were not significantly different (p = 0.30 and p = 0.16, respectively). Swabs detected generic E. coli in 8 of 12 samples, more often than produce leaf samples (3 of 14 positive, Fisher's p = 0.045). Full-length 16S rRNA microbial profiling revealed that swab and produce leaf samples did not show significantly different alpha diversities (p = 0.75) and had many of the most prevalent bacterial taxa in common and in similar abundances. These data suggest that aggregative swabs perform similarly to, if not better than, produce leaf samples in recovering indicators of quality (aerobic and coliform bacteria) and food safety (E. coli), justifying further method development and validation.

Geospatial Data Aggregation Methods for Novel Geographies: Validating Congressional District Life Expectancy Estimates.

Place is a critical determinant of health. Recent novel analyses have explored health outcome estimation for small geographies, such as census tracts, as well as health outcome aggregation to geopolitical geographies with accountable political representatives, such as congressional districts. In one such application, combining these approaches, researchers aggregated census tract estimates of life expectancy to the congressional district level to derive local estimates, but such an approach has not been validated. Here, we compared two sources and approaches to calculating life expectancy data for Pennsylvania congressional districts. We used 2010-2015 census tract life expectancy estimates from the US Small-area Life Expectancy Estimates Project (LEEP) and dasymetric methods to compute population-weighted life expectancy aggregated to the congressional district level. Using georeferenced Vital Statistics data, we aggregated age-specific census tract death and population counts to congressional districts and used abridged life tables to estimate life expectancy. To validate the dasymetric aggregated estimates we compared absolute differences, assessed the correlation, and created Bland-Altman plots to visualize the agreement between the two measures. We found strong agreement between congressional district estimates of life expectancy at birth derived using the dasymetric LEEP model-based approach and Vital Statistics direct estimates approach, though life expectancy at older ages (75 and older) showed weak correlations. This validation contributes to our understanding of geospatial aggregation methods for novel geographies including congressional districts. Health outcome data aggregated to the congressional district geography can support congressional policy making aimed at improving population health outcomes.

Local differential privacy federated learning based on heterogeneous data multi-privacy mechanism

Federated learning enables the development of robust models without accessing users data directly. However, recent studies indicate that federated learning remains vulnerable to privacy leakage. To address this issue, local differential privacy mechanisms have been incorporated into federated learning. Nevertheless, local differential privacy will reduce the availability of data. To explore the balance between privacy budgets and data availability in federated learning, we propose federated learning for clustering hierarchical aggregation with adaptive piecewise mechanisms under multiple privacy-FedAPCA as a way to balance the relationship between privacy preservation and model accuracy. First, we introduce an adaptive piecewise mechanism that dynamically adjusts perturbation intervals based on the data ranges across different layers of the model, ensuring minimized perturbation variance while maintaining the same level of privacy. Second, we propose two dynamic privacy budget allocation methods, which are allocating the privacy budget based on global accuracy and global loss, and allocating the privacy budget based on local accuracy and loss, to ensure that better model accuracy can be achieved under the same privacy budget. Finally, we propose a clustering hierarchical aggregation method in the model aggregation stage, and the model is updated and aggregated after the unbiased estimation of the disturbance in each cluster according to the variance of each layer. FedAPCA improves the balance between privacy preservation and model accuracy. Our experimental results, comparing FedAPCA with the SOTA multi-privacy local differential privacy federated learning frameworks on the MNIST and CIFAR-10 datasets, demonstrate that FedAPCA improves model accuracy by 1%–2%.

Trends in temperature data: Micro-foundations of their nature

Determining whether Global Average Temperature (GAT) is an integrated process of order 1, I(1), or a stationary process around a trend function is crucial for detection, attribution, impact, and forecasting studies of climate change. In this paper, we investigate the nature of trends in GAT building on the analysis of individual temperature grids. Our micro-founded evidence suggests that GAT is stationary around a non-linear deterministic trend in the form of a linear function with one structural break. This break can be attributed to a combination of breaks on individual grids and the standard aggregation method under acceleration in global warming.

Aggregation methods and comparative study in time-to-event analysis models

Aggregation methods and comparative study in time-to-event analysis models

Ensemble Learning-Based Fuzzy Aggregation Functions and Their Application in TSK Neural Networks

AbstractTakagi–Sugeno–Kang fuzzy neural networks (TSKFNN) are powerful tools to model vague or imprecise information. Regression is one of the most important tasks commonly tackled by supervised learning techniques. TSKFNNs are considered suitable models to deal with regression problems on account of their simplicity and flexibility. Aggregation methods play an important role in combining various fuzzy rules from a TSKFNN rule base to obtain a model prediction. However, many current aggregation methods rely on expert experience and human knowledge, which may be hard to acquire and could bring human bias. This paper proposes data-driven aggregation functions for rules aggregation based on ensemble learning, namely AdaBoost and bagging, which can achieve superior generalizability in testing compared with the existing rule aggregation methods. Furthermore, they can also provide insights into the importance of each rule in the model’s decision making, thus, helping to improve the interpretability of the model. Extensive experiments on 11 commonly used benchmark datasets with various sizes and dimensionalities validated the superiority of the proposed ensemble learning-based fuzzy aggregation functions compared with existing state-of-the-art TSKFNNs.

Long-Term Minimum Tillage and Straw Retention Promote Macroaggregate Formation, Carbon and Nitrogen Sequestration under Wheat-Maize Rotation in Northern China

Conservation tillage is believed to promote soil aggregate stability, carbon (C) and nitrogen (N) sequestration, but the underlying mechanisms remain unclear. In this study, soil samples from an 18-year experiment including conventional tillage with straw removal (CT), deep scarification with straw mulching (DS), and no-tillage with straw mulching (NT) were used to obtain different fractions based on a comprehensive wet-sieving method of aggregate and particle size. The results showed that NT and DS increased soil organic carbon (SOC) and N by 9.3–16.4% and 10.8–25.8%, respectively, in addition to increasing the weight proportion of macroaggregates and the contribution of macroaggregate-associated C and N to total SOC and N. The C change in the total POM accounted for 77.4% and 79.9% of the total SOC increase by NT and DS, while the MAOM only accounted for 29.2% and 25.2%, respectively. Meanwhile, microaggregates-within-macroaggregates accounted for 96.9% and 90.5% of the SOC increase by NT and DS, respectively. The total SOC and N were positively correlated with the C and N of the macroaggregates and subfractions. In conclusion, the formation of macroaggregates drives soil C and N sequestration under conservation tillage, and POM and mM were important functional pools in this process.

Whole slide image-based weakly supervised deep learning for predicting major pathological response in non-small cell lung cancer following neoadjuvant chemoimmunotherapy: a multicenter, retrospective, cohort study.

Develop a predictive model utilizing weakly supervised deep learning techniques to accurately forecast major pathological response (MPR) in patients with resectable non-small cell lung cancer (NSCLC) undergoing neoadjuvant chemoimmunotherapy (NICT), by leveraging whole slide images (WSIs). This retrospective study examined pre-treatment WSIs from 186 patients with non-small cell lung cancer (NSCLC), using a weakly supervised learning framework. We employed advanced deep learning architectures, including DenseNet121, ResNet50, and Inception V3, to analyze WSIs on both micro (patch) and macro (slide) levels. The training process incorporated innovative data augmentation and normalization techniques to bolster the robustness of the models. We evaluated the performance of these models against traditional clinical predictors and integrated them with a novel pathomics signature, which was developed using multi-instance learning algorithms that facilitate feature aggregation from patch-level probability distributions. Univariate and multivariable analyses confirmed histology as a statistically significant prognostic factor for MPR (P-value< 0.05). In patch model evaluations, DenseNet121 led in the validation set with an area under the curve (AUC) of 0.656, surpassing ResNet50 (AUC = 0.626) and Inception V3 (AUC = 0.654), and showed strong generalization in external testing (AUC = 0.611). Further evaluation through visual inspection of patch-level data integration into WSIs revealed XGBoost's superior class differentiation and generalization, achieving the highest AUCs of 0.998 in training and robust scores of 0.818 in validation and 0.805 in testing. Integrating pathomics features with clinical data into a nomogram yielded AUC of 0.819 in validation and 0.820 in testing, enhancing discriminative accuracy. Gradient-weighted Class Activation Mapping (Grad-CAM) and feature aggregation methods notably boosted the model's interpretability and feature modeling. The application of weakly supervised deep learning to WSIs offers a powerful tool for predicting MPR in NSCLC patients treated with NICT.

Evaluating Observer Reliability and Diagnostic Accuracy of CT-LEFAT Criteria for Post-Treatment Head and Neck Lymphedema: A Prospective Blinded Comparative Analysis of Oncologist Human Inter-Rater Performance.

Radiation-associated lymphedema and fibrosis (LEF) is a significant toxicity following radiation therapy (RT) for head and neck cancer (HNC) patients. Recently, the CT Lymphedema and Fibrosis Assessment Tool (CT-LEFAT) was developed to standardize LEF diagnosis through fat stranding visualized on CT. This study aims to evaluate the inter-observer reliability and diagnostic accuracy of the CT-LEFAT criteria. This study retrospectively evaluated 26 HNC patients treated with RT that received a minimum of two contrast-enhanced CT scans. Qualitative review was conducted by five physician raters to assess the fat stranding observed on CT according to the CT-LEFAT criteria. Fleiss' kappa analysis was used to assess the inter- and intra-rater reliability, and Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) analysis was used to evaluate diagnostic accuracy. The inter-rater reliability across the six CT-LEFAT regions generally indicated a slight to fair agreement across all raters (0.04 ≤ kappa ≤ 0.36). Intra-observer agreement was generally fair to moderate (overall kappa=0.44). The ROC AUC analysis varied based on aggregation method used (0.60 ≤ average AUC ≤ 0.70). This specific use-case evaluating CT-LEFAT criteria displays limited performance. This suggests that additional materials, such as further training, refinement of imaging methods, or other processes may be required before achieving clinically-ready diagnostic performance of LEF diagnosis.

MAFormer: A cross-channel spatio-temporal feature aggregation method for human action recognition

Human action recognition has been widely used in fields such as human–computer interaction and virtual reality. Despite significant progress, existing approaches still struggle with effectively integrating hierarchical information and processing data beyond a certain frame count. To address these challenges, we introduce the Multi-AxisFormer (MAFormer) model, which is organized in terms of spatial, temporal, and channel dimensions of the action sequence, thereby enhancing the model’s understanding of correlations and intricate structures among and within features. Drawing on the Transformer architecture, we propose the Cross-channel Spatio-temporal Aggregation (CSA) structure for more refined feature extraction and the Multi-Axis Attention (MAA) module for more comprehensive feature aggregation. Moreover, the integration of Rotary Position Embedding (RoPE) boosts the model’s extrapolation and generalization abilities. MAFormer surpasses the known state-of-the-art on multiple skeleton-based action recognition benchmarks with the accuracy of 93.2% on NTU RGB+D 60 cross-subject split, 89.9% on NTU RGB+D 120 cross-subject split, and 97.2% on N-UCLA, offering a novel paradigm for hierarchical modeling in human action recognition.

IGFBP2 functions as an endogenous protector against hepatic steatosis via suppression of the EGFR-STAT3 pathway

ObjectiveNon-alcoholic fatty liver disease (NAFLD) is deemed as an emerging global epidemic, whereas the underlying pathogenic mechanism remains to be clarified. We aimed to systemically analyze all the NAFLD-related gene expression datasets from published human-based studies, by which exploring potential key factors and mechanisms accounting for the pathogenesis of NAFLD. MethodsRobust rank aggregation (RRA) method was used to integrate NAFLD-related gene expression datasets. For fatty liver study, adeno-associated virus (AAV) delivery and genetic knockout mice were used to create IGFBP2 (Insulin-like growth factor binding protein 2) gain- or loss-of function models. Western blot, Co-immunoprecipitation (Co-IP), immunofluorescent (IF) staining, luciferase assay, molecular docking simulation were performed to reveal the IGFBP2-EGFR-STAT3 axis involved. Key axis protein levels in livers from healthy donors and patients with NAFLD were assessed via immunohistochemical staining. ResultsBy using RRA method, the present study identified IGFBP2 being the most significantly down-regulated gene in all NAFLD subjects. The decreased IGFBP2 expression was further confirmed in the liver tissues from patients and animal models of NAFLD. IGFBP2 deficiency aggravated hepatic steatosis and NASH phenotypes and promoted lipogenic gene expression both in vivo and in vitro. Mechanistically, IGFBP2 directly binds to and regulates EGFR, whereas blockage of the IGFBP2-EGFR complex by knockdown of IGFBP2 resulted in the EGFR-STAT3 pathway activation, which in turn promoted the promoter activity of Srebf1. By using molecular docking simulation and protein-protein interaction analysis, the sequence of 233-257 amino acids in IGFBP2 was characterized as a key motif responding for its specific binding to EGFR and the protective effect against hepatic steatosis. ConclusionsThe current study has, for the first time, identified IGFBP2 as a novel protector against hepatosteatosis. The protective effect is mediated by its specific interaction with EGFR and thereby suppressing the EGFR-STAT3 pathway. Therefore, pharmaceutically targeting the IGFBP2-EGFR-STAT3 axis may provide a theoretical basis for for the treatment of NAFLD/NASH and the associated diseases.

Vehicle trajectory prediction method integrating spatiotemporal relationships with hybrid time-step scene interaction

In vehicle trajectory prediction, constructing the interactive relationships among vehicles within the traffic environment poses a significant challenge. Existing models predominantly focus on temporal dependencies within vehicle histories and spatial correlations among neighboring vehicles, overlooking the continuous influence of historical vehicle states on the current time step and the interplay of multiple sequences over time. To address these limitations, we propose a method for multimodal vehicle trajectory prediction that integrates Hybrid Time-step Scene Interaction (HTSI) into the spatiotemporal relationships. Firstly, we introduce the HTSI module, comprising Multi-step Temporal Information Aggregation (MTIA) and Single-step Temporal Information Aggregation (STIA) methods. MTIA utilizes multi-head attention mechanisms to capture temporal dependencies between consecutive frames, thereby generating new time series amalgamating the ongoing influence of historical time states on the current timestamp. Simultaneously, STIA employs multi-head attention mechanisms to capture the spatial dimension weights of multiple time series and, by aggregating spatial interaction features at each timestamp, generates new time series fused with spatial interaction influences. Subsequently, feature extraction is performed through LSTM layers. Moreover, we propose an improved DIPM pooling module, improving the model’s long-term prediction capability by selectively reusing historical hidden states. Ultimately, based on training results from the HighD and NGSIM datasets, our model demonstrates significant advantages in long-term prediction compared to other state-of-the-art trajectory prediction models. Specifically, within the 5 s prediction window, the model achieved a root mean square error (RMSE) of 2.79 m on the NGSIM dataset, representing a 33.62% improvement over the baseline model’s average accuracy. Additionally, on the HighD dataset, the model attained an RMSE of 2.16 m, reflecting a 33.43% enhancement. The crucial code can be obtained from the provided link: https://github.com/gyhhq/Prediction-trajectory .

A Hybrid Feature Selection through Ensemble Rank Aggregation to Improve Perdition and Classification Accuracy

Today, extending credit is a difficult operation since credit databases contain a large quantity of data as well as redundant and useless information. The surplus and unrelated data can lower the categorization and prediction accuracy. In this case, feature selection is crucial for managing massive data. Different rank aggregation (RA) methods are there to ensemble individual feature selection. However not every dataset will perform at its best when using a single RA technique, the combination of multiple RA techniques are needed to increase the perdition accuracy. This study proposes a hybrid rank aggregation model that select features that are significant across different rank aggregation methods like MC4, Kendall tau distance and Borda. This study observed that the performance of ensemble rank aggregation techniques is better than the existing individual rank aggregation methods.

Uncertainty-Aware Federated Reinforcement Learning for Optimizing Accuracy and Energy in Heterogeneous Industrial IoT

The Internet of Things (IoT) technology has revolutionized various industries by allowing data collection, analysis, and decision-making in real time through interconnected devices. However, challenges arise in implementing Federated Learning (FL) in heterogeneous industrial IoT environments, such as maintaining model accuracy with non-Independent and Identically Distributed (non-IID) datasets and straggler IoT devices, ensuring computation and communication efficiency, and addressing weight aggregation issues. In this study, we propose an Uncertainty-Aware Federated Reinforcement Learning (UA-FedRL) method that dynamically selects epochs of individual clients to effectively manage heterogeneous industrial IoT devices and improve accuracy, computation, and communication efficiency. Additionally, we introduce the Predictive Weighted Average Aggregation (PWA) method to tackle weight aggregation issues in heterogeneous industrial IoT scenarios by adjusting the weights of individual models based on their quality. The UA-FedRL addresses the inherent complexities and challenges of implementing FL in heterogeneous industrial IoT environments. Extensive simulations in complex IoT environments demonstrate the superior performance of UA-FedRL on both MNIST and CIFAR-10 datasets compared to other existing approaches in terms of accuracy, communication efficiency, and computation efficiency. The UA-FedRL algorithm attain an accuracy of 96.83% on the MNIST dataset and 62.75% on the CIFAR-10 dataset, despite the presence of 90% straggler IoT devices, attesting to its robust performance and adaptability in different datasets.

An integrated framework for measuring sustainable rural development towards the SDGs

Rural areas are essential for achieving the Sustainable Development Goals (SDGs), necessitating the evaluation of rural sustainability transitions. A precondition for evaluating sustainable rural development is to adapt the SDGs to rural contexts while maintaining global comparability. Besides, tradeoff relationships among evaluation indicators should be captured to overcome the limitations of traditional weighting aggregation methods. In this study, we propose an integrated framework that incorporates the SDGs into a composite indicator system for assessing sustainable rural development. Our framework quantifies these indicators using the index thresholds outlined in the SDGs dashboard, along with customized rules that reflectrural contexts. Furthermore, we employ the Pareto ranking approach to evaluate the sustainable development potentials of rural areas regarding tradeoff relationships among different indicators. In a case study of Zhaoyuan City, China, our findings indicate that only 20 % of the sustainable development goals are achieved on average across all indicator dimensions. The economic and social dimensions are closer to their ideal targets and more compatible with each other, while the ecological dimension remains relatively unsustainable. Sustainable rural development depends primarily on natural resource endowment, ecological sustainability, and residential sustainability of villages. Understanding tradeoff relationships among different indicators enables us to better adapt rural policies to the specific contexts of villages.