Careful monitoring of the quality and quantity of water resources plays a significant role in modern environmental management. Thus, to achieve this aim, the quantitative and qualitative parameters of water resources should be measured and controlled with a desirable accuracy. However, it is not always possible to measure these parameters with easy, inexpensive, precise, and quick experimental methods. Therefore, today to solve this type of problems, new methods including smart methods are used which have a great potential in many computational areas. Considering the variety of the studies and absence of a comprehensive review paper, this research should be conducted. The aim of this paper is to comprehensively review application of the smart methods of artificial neural network and support vector machine in the area of water resources and to develop a comprehensive study source for the researchers interested in this field. The results of these studies all show that these advanced smart methods are more efficient, accurate, economical, and faster than other computational methods to predict quantitative and qualitative parameters of water resources.
This paper shows a Min-Max property existing in the connection weights of the convolutional layers in a neural network structure, i.e., the LeNet. Specifically, the Min-Max property means that, during the back propagation-based training for LeNet, the weights of the convolutional layers will become far away from their centers of intervals, i.e., decreasing to their minimum or increasing to their maximum. From the perspective of uncertainty, we demonstrate that the Min-Max property corresponds to minimizing the fuzziness of the model parameters through a simplified formulation of convolution. It is experimentally confirmed that the model with the Min-Max property has a stronger adversarial robustness, thus this property can be incorporated into the design of loss function. This paper points out a changing tendency of uncertainty in the convolutional layers of LeNet structure, and gives some insights to the interpretability of convolution.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.