Recently, replacement of deep learning with neural network is accelerating in the domains to which neural networks have been applied. And there have been many reports that the deep learning contributed to high accuracy and efficiency. However, it has been pointed out that there are some disadvantages for deep learning. In particular, "the reasons for high accuracy and efficiency are not clear" becomes a critical issue. On the other hand, research to explore the mechanism of deep learning is done theoretically and by detailed analysis of internal state of deep learner. The latter representative method is deep visualization. In this paper, we first explain some fundamentals about deep learning based on neural network. And then, by deep visualization, we show that what has been revealed about the reasons for high accuracy and efficiency, and introduce some examples of use of the intermediate output measures.

In this paper, we pointed out the necessity of low-dimensional feature extraction in the fields of both artificial intelligence and visualization, and outlined a feature extraction method in the field of artificial intelligence, auto-encoder. We firstly described the structure of the auto-encoder and pointed out that the structure is related to a method recently used in the visualization field. In addition, the history that auto-encoder contributed to the development of present artificial intelligence is described, and recent research trend of auto-encoder is introduced. Finally, as an application example in visualization, we described our research using an auto-encoder for visualization of temporal behavior of an unsteady flow. It is expected that new usage of the structure of auto encoder will develop in both fields in the future.

With the advances in supercomputers and observation instruments, “Climate Big Data” including numerical simulation and observational data have been generated and accumulated. A convolutional neural network, which is one of deep neural networks used for image pattern recognition, is applied for analyzing climate big data. In the present manuscript, we introduce (i) detection of precursors of tropical cyclone from climate simulation data, (ii) classification of cloud type, and (iii) estimation of cloud coverage from cloud image and visualization image of simulation data.

This article describes a trend and usage of AI, especially, deep learning in medical area. AIs including Machine learning methods are used in this medical area, and deep learning is attractive for many researchers as well as other applications because of high precision. This article reviews architectures and usages of deep learning for medical images. Also, we discusses some issues on the usage of medical images with deep learning in visualization areas.

In recent years, as the number of scientific articles submitted to refereed journals has increased, the burden of peer review of researchers is increasing. The increase in peer review burden has led to delays in the publication of articles and deterioration in the quality of peer-review, and the collapse of the peer-reviewed system that has supported science is also a concern. Therefore, it is necessary to develop support technology for peer review to reduce the burden of researchers. In this research, we consider peer-review support using artificial intelligence technology. We used a Doc2Vec to numerically process the text structure of the scientific articles. We showed the differences in the text structure of the accepted and rejected manuscripts of 591 abstracts submitted to the Journal of Visualization. Furthermore, we developed a classification model of acceptance and rejection of the articles using SVM. We achieved a classification accuracy of 75% only with the abstract of the articles.

A size and complexity of digital games increase recently. Even for game developers, they cannot understand a whole image of a game which they are developing. But AI can support game developers by collecting data and analyze them to abstract important features and visualize them to them. In this article, by using FINAL FANTASY XV cases, it is showed how AI can support game developers by visualization of game data.