Historical Evolution and Future Works in Artificial Neural Network (ANN) Technological Development

― Using ANN-related studies in Japanese hydrologic and hydraulic engineering ―

Abstract

This paper describes the historical evolution of artificial neural networks (ANN) technology from the initial ANN to the applied ANN, including deep learning. The paper also explains key studies associated with ANN technology in Japanese hydrologic and hydraulic engineering fields during recent 30 years. The key studies were selected from the peer-reviewed, water-related journals, such as the Japan Society of Civil Engineers, the Japan Society of Hydrology and Water Resources, and the Japanese Society of Irrigation, Drainage and Rural Engineering. We have pointed out the merits and demerits of ANN application to the water-related studies and have discussed the reason why deep learning has recently become popular in their fields. In addition, we have proposed potential future works of ANN based on valuable knowledge of past and latest studies.