In this study, we question the value of social robots in hotel rooms to provide interactive services in a humanoid form. We conducted a field experiment in which a mechanoid social robot (smart speaker) or a humanoid social robot provided interactive services in a hotel room, and conducted a comparative evaluation. The results show that humanoid social robots increase the use of interactive services in hotel rooms and improve hotel guests’ impressions of interactive services. In addition, there was no significant difference in privacy concerns between mechanoid and humanoid social robots, suggesting that humanoid form are not an obstacle to the use of the service. These results indicate the value of the hotel industry’s incorporation of humanoid social robots into room services.

Different individual faces different difficulties during language development. Especially, when phonological awareness, which intentionally makes individuals aware of phonological units, is immature, diverse speech errors occur. For dealing with such diversity, it is important to provide support based on an understanding of the internal mechanisms of learners. In this article, we propose a system to support phonological awareness formation and confirm the system’s feasibility in the adaptation process to individual error patterns. The proposed system subjects a word-choice-based Shiritori game as a task and implements cognitive models to represent the individual’s state of phonological awareness. Based on the user’s repeated choices of words proposed by the model, the system eventually selects a valid model for estimating the user’s state. As a result of an experiment using the proposed system, we confirmed that participants’ preference for models was varied by phonological problems artificially generated by audio filters in a laboratory. This result supports the assumption that repetitive interactions with cognitive models can estimate individual problems of phonological awareness.

In this study, we propose an evaluation system that utilizes natural language processing to detect stumbling points among individuals learning dementia care, with the aim of contributing to the development of expertise in dementia care education. Thirteen skilled occupational therapists and eleven occupational therapy students organized clinical pictures of three virtual dementia cases based on the ICF framework. Utilizing the descriptions obtained and corresponding ICF codes for training, we screened for 'deficit descriptions' and 'extra descriptions' among novices. The identification of features among novices, such as a lack of focus on residual functions in dementia patients and insufficient recognition of the importance of social connections, proved to be incredibly valuable. Capturing these stumbling points of novices forms the foundation for interface design connecting learners and educators, aiming to optimize both and achieve an enhanced learning support environment, with an expectation for improving the quality of dementia care at the field level. Additionally, the scheme of detecting stumbling points using the ICF and extracting the thinking features of professionals demonstrated its potential applicability to the existing educational content.

Despite recent successes in nonverbal human-computer interaction (HCI) facilitated by deep learning methods, sign language translation for HCI remains underexplored. In this paper, we analyze and develop a sign language translation system that can recognize continuous signs and convert the sign meanings into natural spoken sentences in an end-to-end manner. We believe this system will enhance the interaction between computers and “deaf and hard-of-hearing individuals”. In developing this sign language translation system, we introduced high-quality sign embedding to extract informative spatial-temporal representation from continuous sign motions and adopted label smoothing to the training criteria to mitigate the overfitting issue. The proposed methods, therefore, help narrow the modality gap between vision (sign language) and language (spoken sentence). We conducted the experiments with the proposed methods on the PHOENIX14T dataset, yielding significantly improved results (WER↓: 29.20→20.79, BLEU-4↑: 21.12→ 24.56).

The process of manipulating the image that others have of oneself by some means is called self-presentation. Self-presentation in real space has been based on the assumption that one’s “one and only body” is the basis of self-presentation, but in virtual space, it is possible to completely free oneself from the constraints of reality by using an avatar as a substitute for oneself. The purpose of this study is to clarify what strategies are used in the creation of avatars, and to what extent the self-presentation image held in real space is projected onto the avatar. We conducted an experiment in which 30 participants (14 males and 16 females) were asked to create their avatars. The results revealed that the three central strategies were the ideal self-actualization strategy, which involves creating an ideal avatar; the reality self-projection strategy, which involves projecting the real self; and the likability acquisition strategy, which involves creating an avatar that others will like. We also found that the self-presentation images held in real space were also projected onto avatars, and were found to be influenced by personal attributes in real space, such as gender and desire for approval. The avatars created by the task setting of this study can be considered entry models for self-avatars and are expected to contribute in the future to psychological counseling situations and support situations related to identity formation and development.

For videos of oral presentations using visual materials such as photographs and diagrams, it is known that the viewers' comprehension is improved if the presenters appear in the frame with gestures. However, the necessity of gestures should depend on the linguistic expressions used with those gestures, and it is not clear whether the viewers actually pay attention to the gestures regardless of their necessity. This paper describes an early result of our gaze analysis for the viewers watching videos in which an avatar describes a route with linguistic expressions and gestures by showing a map.

As represented by automated driving, with the development of AI technology and automation, multitasking situations have increased where people can monitor automated tasks and perform other tasks concurrently. This study investigated whether the attentional function can be regulated in multitasking, including visual monitoring tasks by directing the proportion of attentional allocation. As a result, we found that the performances of two tasks can be increased or decreased according to verbal instructions, while overall multitasking performance remained unaffected.

Feedback can be in the form of praise or caution. However, the impacts of these variations on the effectiveness of feedback, particularly regarding smartphone overuse, remain unexplored. Furthermore, the difference in underlying motivating mechanisms arising from the types of feedback have not been thoroughly examined. Therefore, this study investigates the differential effects of praise and cautionary feedback, as well as feedback on used and endured periods. We conducted a series of experiments, employing subjective assessments to measure feedback's influence on motivation, with a participant pool of 12 university students. Our analysis of variance, considering praise/cautionary feedback and used/endured periods, indicates that praise feedback enhances motivation more effectively than cautionary feedback, while no significant differentiation was observed between feedback regarding usage and endurance. Furthermore, our path analysis integrating subfactors unveils a mechanism by which praise feedback augments motivation by eliciting positive emotions, whereas cautionary feedback enhances motivation through the promotion of acceptance.

Sign Language Generation (SLG) is one of the emerging fields in Sign Language (SL) communication technologies. SLG employs a Sequence to Sequence (Seq2Seq) model with SL glosses serving as intermediaries. While using Seq2Seq is an optimal way of generating SL, it overlooks the fineness and smoothness of the inbetween transitions for SL glosses. In this work, we address the issues of fine motion prediction and smooth transition generation to seamlessly integrate glosses for natural SL expression in SLG termed as Sign Language Inbetweening (SLI). In SLI, we propose the Body Parts Integration (BPI) for fine motion prediction and the Variational Autoencoder-Generative Adversarial Network (VAE-GAN) generator with a progressive embedding for smooth transition generation over variational transition lengths. Evaluation of the proposed BPI-VAE-GAN model is conducted on a Motion Capture dataset of Japanese Sign Language performed by a native signer. Assessment employs Mean Angular Error (MAE) for fine motion prediction and mean Jerk for smooth transition generation. Results indicate that the proposed architecture significantly reduces MAE, outperforming the GAN by 46% reduction and considerably outperforms the linear interpolation. Additionally, Jerk analysis illustrates the system’s capacity to generate smoother motion sequences compared to GAN.

We used Puyo-puyo esports, one of the computer game software, and focused on measuring movement-related cortical potential (MRCP) and heart-beat evoked potential (HEP) as brain activities to evaluate the degree of acquiring the proficiency of the Puyo-puyo esports. To clarify the modulations of the MRCP and the HEP to parameters related to the experiment, we constructed 64 statistical models and performed model selection based on Akaike information criteria for the models. For the MRCP, a model that included day was selected. The MRCP increased over the course of the experimental days, indicating that participants gradually increased the amount of brain resources in the process of mastering Puyo-puyo esports. For the HEP at the fronto-central electrodes, a model that includes days was selected. The HEP at the fronto-central electrodes deflected in a positive direction as the experimental day progressed, suggesting that the participants shifted their attention to exoceptive perception to interoceptive perception as the proficiency of Puyo-puyo increased. We showed the possibility that MRCP and HEP can evaluate the process of acquiring the proficiency of the computer game.

Curling irons can create curled hair by heating it. However, because the operation of the curling iron involves six degrees of freedom (6DoF) movements, such as translation and rotation, it is difficult to understand and reproduce by simply watching instructional videos. In this study, we proposed Color-Path, a smart mirror-shaped system that allows users to easily understand how to move a curling iron. The proposed system acquires the moving paths of the curling iron using a camera on a smart mirror and an accelerometer on the curling iron. The system displays the curling iron paths of the target hairstyle on a smart mirror in Augmented Reality (AR). From the evaluation experiments, it was confirmed that the curling irons contributed to reducing the difference from the target in the relative coordinates to the face position and posture of the curling irons through quantitative evaluation. However, a subjective evaluation showed that our system could not reproduce the appearance of the target hairstyle. The results indicate that increasing the density of presentations on the moving paths is important. They also indicate that the time to heat the hair should be considered.