In this paper, we describe possible applications of early exit deep neuralnetworks in magnetic resonance imaging, aiming to improve patient scan times and reduce processing costs. The solutions rely on deep neural ne...
详细信息
Privacy preservation is critical for neuralnetwork inference, which often involves collaborative execution of different parties to make predictions on sensitive data based on sensitive neuralnetwork models. However,...
详细信息
Internet of Things (IoT)-enabled Smart Energy Management (SEM) in distributed Energy Resources (DERs), while crucial for optimizing energy distribution and resource management faces challenges such as data inconsisten...
详细信息
ISBN:
(纸本)9798331523923
Internet of Things (IoT)-enabled Smart Energy Management (SEM) in distributed Energy Resources (DERs), while crucial for optimizing energy distribution and resource management faces challenges such as data inconsistencies and the variability of renewable energy generation. These issues result in inaccurate demand forecasts, leading to suboptimal allocation of resources and inefficient Energy Management (EM). Additionally, as energy demand and generation patterns are influenced by factors like weather conditions, time of day, and energy consumption behaviors, prediction models may struggle to capture these dynamic changes, affecting the reliability of forecasts. To address these limitations, this manuscript proposes a novel approach for energy demand prediction. Data is collected from IoT-enabled sensors monitoring DERs. The data undergoes pre-processing, where the Fast Resampled Iterative Filtering (FRIF) method is used to eliminate missing values and normalize the inputs. The Self-Adaptive Physics-Informed neuralnetwork (SAPINN) model then utilizes the processed data to forecast energy demand, renewable energy generation, and storage levels. Green Anaconda Optimization (GAO) is applied to optimize the weight parameters of the SAPINN model. The proposed SAPINN-GAO method is implemented using the MATLAB platform and compared with existing models, such Stacked Convoluted Bi-Directional Gated Attention network-Hybrid Darts Seagull Optimizer (SConBGAN-HDSO), Recurrent neuralnetwork (RNN), and Support Vector Machine-Particle Swarm Optimization (SVM-PSO). The SAPINN-GAO method achieves an accuracy of 99.2%, precision of 99.2%, and a Root Mean Square Error (RMSE) of 2.2%, demonstrating its superior performance in energy demand prediction. The SAPINN-GAO method's higher accuracy and precision, coupled with its robust performance, make it a reliable and efficient solution for energy demand, renewable energy generation, and storage levels forecasting in IoT-enabled SEM syst
This paper presents a new approach to the hyper-distributed hyper-parallel implementation of the artificial intelligent (AI) heuristic algorithms for real-time searching, matching and planning. By using the competitiv...
详细信息
This paper presents a new approach to the hyper-distributed hyper-parallel implementation of the artificial intelligent (AI) heuristic algorithms for real-time searching, matching and planning. By using the competitive activation mechanism of dynamically clustering neuralnetworks, the concurrent propagations and competitions of concurrent autowaves yielded by distributed parallel heuristic AI algorithms for searching any implicit AND/OR graph are realized. Compared with the AI approaches based on the conventional sequential symbolic logic and the conventional neuralnetworks, the approach of this paper has many advantages in many respects, such as high processing speed, always successful obtainment of the optimal solution, local connections between cells, easy utilization of heuristic knowledge, and feasibility of the VLSI implementation.
In face of constant fluctuations and sudden bursts of data stream, elasticity of distributed stream processing system has become increasingly important. The proactive policy offers a powerful means to realize the effe...
详细信息
For ECG signal processing, information extraction from a noisy background is the fundamental objective. Filtering (noise suppression, baseline wander elimination) is a very important step in efficient ECG signal featu...
详细信息
ISBN:
(纸本)142440360X
For ECG signal processing, information extraction from a noisy background is the fundamental objective. Filtering (noise suppression, baseline wander elimination) is a very important step in efficient ECG signal features extracting to enhance the performance of automatic detection and classification of different cardiac diseases. In this paper we used distributed Approximating Functional (DAF) wavelets to develop algorithms for signal approximation and filtering. These algorithms use Moving Average Artificial neuralnetwork with Wavelet type Hermite activating function. They are evaluated in MATLAB with signals from the MIT-BIH arrhythmia database and comparisons are made with the classical (radial basis function and sigmoid type activating function) artificial neuronal networks. New functions were created and integrated into MATLAB environment. The outcomes indicate a good performance tradeoff between accuracy and response time, making this type of algorithms desirable also for real-time implementation.
Scalable data management is essential for processing large scientific dataset on HPC platforms for distributed deep learning. In-memory distributed storage is preferred for its speed, enabling rapid, random, and frequ...
详细信息
ISBN:
(纸本)9798350355543
Scalable data management is essential for processing large scientific dataset on HPC platforms for distributed deep learning. In-memory distributed storage is preferred for its speed, enabling rapid, random, and frequent data access required by stochastic optimizers. Processes use one-sided or collective communication to fetch remote data, with optimal performance depending on (i) dataset characteristics, (ii) training scale, and (iii) interconnection network. Empirical analysis shows collective communication excels with larger mini-batch sizes and/or fewer processes, whereas one-sided communication outperforms at larger scales. We propose MDLoader, a hybrid in-memory data loader for distributed graph neuralnetwork training. MDLoader features a model-driven performance estimator that dynamically selects between one-sided and collective communication at the beginning of training using Tree of Parzen Estimators (TPE). Evaluations on NERSC Perlmutter and OLCF Summit show MDLoader outperforms single-backend loaders by up to 2.83× and predicts the suitable communication method with 96.3% (Perlmutter) and 94.3% (Summit) success rate.
distributed DNN inference is becoming increasingly important as the demand for intelligent services at the network edge grows. By leveraging the power of distributed computing, edge devices can perform complicated and...
详细信息
Abstract: he possibility of neuralnetwork signal processing of a tomography quasi-distributed fiber-optic measuring network is shown. The measuring network presents several nonlinear measuring lines with fiber-optica...
详细信息
暂无评论