In this paper, we propose an attribute added face image generation system using Deep Convolutional Generative Adversarial Networks (DCGANs). Convolutional Neural Networks (CNNs) can extract important features of an image and attain high precision in image classification tasks. In the proposed system, image features are extracted using CNNs, and attribute features added to image features, and attributes added images are generated by DCGANs. Specifically, we use the attributes of “smile” and “male”, and work on a task of generating smile images from non-smile images, and a task of generating male images from women images. Since the training of the proposed system requires image pairs including with and without attributes, we use two extraction methods using attribute label and cosine similarity. Attribute features are defined as the averaged difference between image features with and without attributes. We performed two kinds of evaluation experiments, and excellent characteristics were obtained.
This paper uses Hierarchical Fuzzy Integrals (HFI) for comprehensive evaluation models with interaction among evaluation criteria and an extended version of the Analytical Hierarchy Process (AHP) model. First, we show the structure and computational procedure of HFI models on a single-level hierarchy diagram. A sensitivity analysis with interaction degree shows that the selected facility is different depending on the interaction degree. Second, we extend the multi-level hierarchical model and present its computational procedure. We select the model that includes Kansei evaluation items and propose the standard hierarchical relationship and standard interaction degrees using the facility selection model for the elderly. Lastly, we provide some remarks for identifying individual evaluation scores.
The authors are conducting research on value creative consensus building process. This study takes in case of multiple-choice as an example of value creative consensus building process. The authors observed the phenomenon that viewpoint has changed by sharing the conception (value). Next, the authors analyzed the process using constructed Bayesian network model and showed that the appearance of the conception influences the selection. As described above, the authors were able to explain quantitatively the structure of consensus building process that has many choices using mathematical method called by Bayesian network. Furthermore, in this study, we compared consensus building processes based on quantity of choices. This shows that the consensus building process can be grasped by the same structure irrespective of the number of choices, and the possibility that various consensus building processes can be discussed by quantitative analysis in the same way.
In a non-task oriented dialogue system, it is important that the system can expand various topics for continuing a conversation. At the conventional dialogue systems, topic expansion has not been paid attention. The conventional dialogue systems often use the user's input topic word as a topic word of system's utterance. In this study, we propose a topic word expansion method. Concretely, we extract a topic word from user's utterance and generate embeddings using genetic algorithm and embeddings of the topic word from the user's utterance. As the evaluation function of genetic algorithm, we use neural network that learns topic word correspondence in a dialogue corpus, lexical knowledge and heuristic rules. By using the proposed system, we aim to model ambiguity of topic words and make it possible to expand topic words of conversation integrating plural knowledge. As evaluation experiments, we evaluated the proposed system by subjectivity and verified the effectiveness.