Miniaturization of recording devices for electroencephalogram (EEG) is important to develop wearable EEG devices. In recent years, low-cost wearable devices for EEG monitoring have been developed and are commercially available. Daily health management and early detection of brain disease might be possible using wearable EEG devices. However, reports of research or clinical applications of wearable EEG devices are few. Therefore, the aim of this study was to investigate whether wearable EEG devices could detect changes in brain activity. In this study, we determined the association between EEG and emotional image presentation. EEG spectra were analyzed in the resting state and during emotional image presentation. The spectrum values decreased more during positive and negative image presentation than that during the resting state suggesting that decrease in the alpha (α)-wave spectrum may be induced by event-related desynchronization of the α-waves associated with emotional image presentation. In addition, Pearson product-moment correlation coefficient of the α-waves was calculated using eight electrodes to determine the α-wave coupling strength of the entire head. The correlation value between the electrodes of the occipital visual cortex during image presentation tended to be high. However, that of the frontal and occipital areas tended to be low. Thus, α-wave correlation value between the occipital visual cortex was higher than that of other regions during image presentation. These results suggest that wearable EEG devices could detect changes in brain activity associated with attention.

Deep learning solves many classification problems. However, it is difficult to solve problems with imbalanced data. Therefore, the data volume is increased for the purpose of balancing. This is called data augmentation. Generally, the method of image data augmentation uses noise addition, rotation, and the like. Recently, images are generated using the generative adversary network: GAN. However, data augmentation methods are difficult in natural language processing. In addition, manual data augmentation is burdensome and requires mechanical methods. Mechanical text augmentation is more difficult than images. Because it is difficult to analyze the feature of sentences. This paper proposes a sentence generation method by machine learning focusing on graph information. The graph information obtained by CaboCha is processed by graph Convolution. The proposed GAN was used to generate sentences, and then three experiments were performed to evaluate its effectiveness.

Fuzzy co-clustering induced by multinomial mixture models (FCCMM) is an effective method for analyzing such cooccurrence information data as document-keyword frequencies, but often suffers from the cluster validation problem due to a priori selection of cluster numbers. In this paper, a modified model of robust cluster number selection in Gaussian mixture models is proposed, where the optimal number of clusters are automatically extracted in FCCMM through rejection of unnecessary clusters considering a novel penalty on cluster volumes.

In human robot interaction (HRI), social robots are expected to equip with social skills to interact with people under diverse situations. The way human perceive robot is a crucial factor to influence the acceptability of robot. To investigate influential feature on robot behavior and understand human preference, this paper briefly presented robot impression analysis of cute and cool style greeting respectively. We modified the behavior on Pepper robot and recruited participants to evaluate the robot expression style. The results showed the relation between expression patterns and tendency which indicated a few groups of preference similarity in small sample analysis. Further classification of more impression preference group is highly anticipated. Undoubtedly, it is essential to personalize robot behavioral styles to satisfy individual necessity and maintain human robot communication.

In this paper, representations of a situation evaluation model and their learning methods in RoboCup soccer simulation 2D are reviewed. A soccer game is commonly known as an example of multi-agent systems in an uncertain and dynamic environment. First, this paper discusses the similarities and differences between the RoboCup soccer simulation and other benchmark tests of game AI research. Next, a search method based on action chains is presented as a mechanism of action selection by a soccer player. This method is based on the look-ahead of field situations by the combination of a search tree and a situation evaluation model, which is the case in the chess and shogi programs. Then, various methods of constructing the situation evaluation models that are used in the generation of action chains by the machine learning framework are reviewed along with example models that have been proposed.