No abstract

No abstract

No abstract

In this research paper, we focused on the developing a secure and efficient time-series forecasting of nuclear reactions using swarm intelligence (SI) algorithm. Nuclear radioactive management and efficient time series for casting of nuclear reactions is a problem to be addressed if nuclear power is to deliver a major part of our energy consumption. This problem explains how SI processing techniques can be used to automate accurate nuclear reaction forecasting. The goal of the study was to use swarm analysis to understand patterns and reactions in the dataset while forecasting nuclear reactions using swarm intelligence. The results obtained by training the SI algorithm for longer periods of time for predicting the efficient time series events of nuclear reactions with 94.58 percent accuracy, which is higher than the deep convolution neural networks (DCNNs) 93% accuracy for all predictions, such as the number of active reactions, to see how the results can improve. Our earliest research focused on determining the best settings and preprocessing for working with a certain nuclear reaction, such as fusion and fusion task: forecasting the time series as the reactions took 0-500 ticks being trained on 300 epochs

Recent years have witnessed the success of artificial intelligence–based automated systems that use deep learning, especially recurrent neural network-based models, on many natural language processing problems, including machine translation and question answering. Besides, recurrent neural networks and their variations have been extensively studied with respect to several graph problems and have shown preliminary success. Despite these successes, recurrent neural network -based models continue to suffer from several major drawbacks. First, they can only consume sequential data; thus, linearization is required to serialize input graphs, resulting in the loss of important structural information. In particular, graph nodes that are originally located closely to each other can be very far away after linearization, and this introduces great challenges for recurrent neural networks to model their relation. Second, the serialization results are usually very long, so it takes a long time for recurrent neural networks to encode them. In the methodology of this study, we made the resulting graphs more densely connected so that more useful facts could be inferred, and the problem of graphical natural language processing could be easily decoded with graph recurrent neural network. As a result, the performances with single-typed edges were significantly better than the Local baseline, whereas the combination of all types of edges achieved a much better accuracy than just that of the Local using recurrent neural network. In this paper, we propose a novel graph neural network, named graph recurrent network.