Recent advances in deep learning have shown many successful stories in smart healthcare applications with data-driven insight into improving clinical institutions’ quality of care. Excellent deep learning models are heavily data-driven. The more data trained, the more robust and more generalizable the performance of the deep learning model. However, pooling the medical data into centralized storage to train a robust deep learning model faces privacy, ownership, and strict regulation challenges. Federated learning resolves the previous challenges with a shared global deep learning model using a central aggregator server. At the same time, patient data remain with the local party, maintaining data anonymity and security. In this study, first, we provide a comprehensive, up-to-date review of research employing federated learning in healthcare applications. Second, we evaluate a set of recent challenges from a data-centric perspective in federated learning, such as data partitioning characteristics, data distributions, data protection mechanisms, and benchmark datasets. Finally, we point out several potential challenges and future research directions in healthcare applications.

Early prediction of students’ learning performance and analysis of student behavior in a virtual learning environment (VLE) are crucial to minimize the high failure rate in online courses during the COVID-19 pandemic. Nevertheless, traditional machine learning models fail to predict student performance in the early weeks due to the lack of students’ activities’ data in a week-wise timely manner (i.e., spatiotemporal feature issues). Furthermore, the imbalanced data distribution in the VLE impacts the prediction model performance. Thus, there are severe challenges in handling spatiotemporal features, imbalanced data sets, and a lack of explainability for enhancing the confidence of the prediction system. Therefore, an intelligent framework for explainable student performance prediction (ESPP) is proposed in this study in order to provide the interpretability of the prediction results. First, this framework utilized a time-series weekly student activity data set and dealt with the VLE imbalanced data distribution using a hybrid data sampling method. Then, a combination of convolutional neural network (CNN) and long short-term memory (LSTM) was employed to extract the spatiotemporal features and develop the early prediction deep learning (DL) model. Finally, the DL model was explained by visualizing and analyzing typical predictions, students’ activities’ maps, and feature importance. The numerical results of cross-validation showed that the proposed new DL model (i.e., the combined CNN-LSTM and ConvLSTM), in the early prediction cases, performed better than the baseline models of LSTM, support vector machine (SVM), and logistic regression (LR) models.

No abstract

The sparse data in PM2.5 air quality monitoring systems is frequently happened on large-scale smart city sensing applications, which is collected via massive sensors. Moreover, it could be affected by inefficient node deployment, insufficient communication, and fragmented records, which is the main challenge of the high-resolution prediction system. In addition, data privacy in the existing centralized air quality prediction system cannot be ensured because the data which are mined from end sensory nodes constantly exposed to the network. Therefore, this paper proposes a novel edge computing framework, named Federated Compressed Learning (FCL), which provides efficient data generation while ensuring data privacy for PM2.5 predictions in the application of smart city sensing. The proposed scheme inherits the basic ideas of the compression technique, regional joint learning, and considers a secure data exchange. Thus, it could reduce the data quantity while preserving data privacy. This study would like to develop a green energy-based wireless sensing network system by using FCL edge computing framework. It is also one of key technologies of software and hardware co-design for reconfigurable and customized sensing devices application. Consequently, the prototypes are developed in order to validate the performances of the proposed framework. The results show that the data consumption is reduced by more than 95% with an error rate below 5%. Finally, the prediction results based on the FCL will generate slightly lower accuracy compared with centralized training. However, the data could be heavily compacted and securely transmitted in WSNs.

No abstract

The National Electricity Company (PT PLN) should have an estimated peak load of the substation transformer in the future. This is useful to be able to achieve transformer capability and can be used as a first step to anticipate the possibility of replacement of a new transformer. This research presents a peak load forecasting system transformer1 and transformer2 in Bumiayu substation using Backpropagation Artificial Neural Network (ANN). This study includes the procedures for establishing a network model and manufacture forecasting system based GUI (Graphic User Interface) using MATLAB 2015a. The formation of the network model refers to input variables consisting of GRDP data, population data and historical data of peak load of transformer. In this research, a multilayer network model, which consists of 1 input layer, 2 hidden layers and 1 output layer, is used. The peak load forecasting of transformer1 produces 5.7593e-08 for training MSE and 5.3784e-04 for testing MSE. Meanwhile, forecasting the peak load transformer2 generated 3.3433e-08 for training MSE and 9,4710e-04 for testing MSE.

No abstract

Development of waste-to-energy plant is regarded as one of a solution to reduce fossil energy dependency and switch to another alternative energy sources, thus needs to be implemented immediately. Research on the potential of organic waste by taking case studies at UMY green campus program can be an illustration that in the future UMY can process this potential into an alternative renewable energy. This research was conducted by studying waste management and data collection methods, such as field observations, discussions with related parties about the waste cycle, and calculations with the concept of gasification. Calculation of biomass produced by organic waste is then supplied to the motor generator for electricity production. Annually, 31.8 tons of organic waste could produce 7.9 kiloliters of biomass and channel it to the power plant system for generating electricity energy as much as 5.3 MWh per year. This number equals to the electricity supply of five small-housing in Indonesian typical house.