This research uncovers the personalization of emotionally and culturally-induced health marketing content, targeting three countries with contrasting cultures according to the culture map model. It aims to demonstrate how tailored content, crafted with narrative techniques aligned with cross-cultural comprehension, enhances persuasive efficacy to influence user behavior. The experimental phase involved the presentation of 19 animated video advertisements with avatars' emotional facial expressions and culturally elicited narratives to 190 participants from the U.S., Japan, and Germany, in a counterbalanced way. As a result, the majority of participants were influenced by emotions evoking happiness, positive surprise, and sadness, with susceptibility primarily associated with traits such as openness, agreeableness, and conscientiousness. The comprehensive investigation that integrates six fundamental emotional components, Erin Meyer culture map, the Big Five personality traits theory, and Fogg's persuasive technology framework highlights the originality of this research.

Generative art is algorithmically generated art that is synthesized by computer software algorithms or mathematical autonomous processes. The learner wears HoloLens2 and learns by transforming and observing three-dimensional (3D) generative art projected on a virtual space. In this study, we developed an experiential learning system for learning mathematics and programming by experiencing generative art in a virtual space. We conducted a comparative study to see if this system would make a difference in learning efficiency compared to learning with a conventional two-dimensional (2D) learning medium such as textbooks. The results showed that there was no significant difference in the learning efficiency between conventional learning and learning using the generative art displayed in the virtual space in terms of learning the volume and surface area of the torus body, however, there was a difference in the way the subjects drew the torus body. We deduced that HoloLens2 will be an effective educational supplementary tool to 3D visualize 3D figures while giving a programming insight through reflections in code obtained from changing the size of the 3D figure.

The number of opportunities for activities that can be done from the comfort of one's home has increased considerably, and recently, activities such as events in the metaverse using virtual avatars have become more accessible. We believe that one of the main reasons for this is the high degree of anonymity of social networks which has led to a state of immersion. Unfortunately, problems may occur in a highly anonymous metaverse. In this experiment, we investi-gate the effect of immersion on communication by changing the anonymity of the person with whom one converses using a virtual avatar in the metaverse. The subjects were di-vided into four groups: anonymous-anonymous, anonymous-non-anonymous, non-anonymous-anonymous, and non-anonymous-non-anonymous. The results of the experiment showed that there were no significant differences between the groups, suggesting that the subjects' anonymity did not affect their immersive communication. We deduce that primitive stages of visual anonymity inhibit rude and violent behaviors in the metaverse. Certainly, a superficial status of primitive visual anonymity could be utilized for different applications. Nevertheless, high levels of anonymity are required to allow de-individualization to take effect in the metaverse.

Social robots are taking major roles in our lives through many sectors, ranging from eHealth to industry. However, social robots that can interact with people are expensive. In this study, we evaluate what makes an inexpensive retro-projected Face-Display/robot different from the advanced social robot “Furhat”. How will a retro-projected robot be perceived compared to an advanced social robot? We built a retro-projected robot and compared it with Furhat. Seventy-two participants interacted with both robots in a between-subjects study design. We compare the two robots using different HRI questionnaires. Furthermore, we investigate which robot face is preferred to interact with, a human face (due to social agency) or an anime face (due to Japanese anime culture). In a retro-projected robot, a human face is preferable as it is rated higher in terms of anthropomorphization, trust, competence, and empathy. An anime face was rated higher in warmth, attractiveness, pleasantness, and comfort to see. In Furhat, few differences were noticed between the human and anime faces. A human face was perceived to be more emphatically responsive, and an anime face was rated highly in trustworthiness, warmth, competence, and empathic understanding. Furhat excelled significantly in the aspects of anthropomorphization, likeability, trustworthiness, warmth, competence, empathic responsiveness, attractiveness, and pleasantness to see. On the contrary, the retro-projected robot excelled in inducing social discomfort. Interestingly, both robots were perceived similarly in aspects of perceived intelli-gence and safety, empathic understanding, and comfort to see. An anime face on either robot is significantly more pleasant and is always perceived to be young in age. We deduce that retro-projected robots are inferior to Furhat in most HRI aspects and similar in a few of them. Thus, retro-projected robots can serve effectively in fields where the three aforementioned indifferent aspects are the main requirements in an application. Moreover, whenever aesthetics and visual pleasantness are needed, an anime face is preferred over a human one.

Embarrassment is a social emotion that stands out due to its relatively short duration. It depends on being witnessed by others where the relationship between the actor and the witness is not close. Surprisingly, despite its importance and relevance, very few studies investigated the embarrassment emotion psychophysiologically. Certainly, a comprehensive study of the embarrassment emotion will benefit the affective computing community and push integrating such emotion into our technological affective systems forward. In our study, we compare the psychophysiological correlates of embarrassment emotion between shy males and females. We collected the electrodermal activity (EDA) and electrocardiogram (ECG) data obtained during experimentally induced embarrassment from 9 shy male subjects. We compare our data to the results of another work where the ECG and EDA data are also collected but for shy female subjects. Our goal is to highlight gender differences in the psychophysiological markers of the embarrassment emotion. Electrodermal behavior was similar and there were no gender differences. Interestingly, gender differences surfaced when ECG data was being studied. Unlike shy females, shy males' heart rate decreases when they are embarrassed. We suspect that the heart rate differences could be the reason behind blushing being more pervasive and prominent in females than males.

When developing commercial games, the difficulty level is carefully designed to satisfy the majority of game players. An inappropriate difficulty level can cause stress for game players and lead them to quit the game. Designing a game that is suitable and enjoyable for everyone is a daunting task, especially with the specific individual needs of everyone. Our goal is to provide insights into game design that can be suitable for males and females. Thus, we investigate gender differences in the game Tetris. We compare subjective ratings, questionnaire responses, electrodermal activity (EDA), and electrocardiogram (ECG) data obtained from 5 males and 5 females while playing Tetris. Our results show that a high feeling of control correlates with high skin conductance level (SCL), skin conductance response (SCR), and low heart rate (HR) which was found in the male sample. Low feelings of control and high anxiety correlate with high HR and low SCL and SCR, which was realized in the female sample. Game self-skill, playing hours, and game experience questionnaire (G EQ) responses do not show gender differences in gameplay. Gender differences were exclusively pronounced when physiological data was studied.

Educational robots are becoming increasingly incorporated into classrooms to teach many subjects including language, science, and also substitute teachers. In recent years, ChatGPT has become easy to use for education but there are problems if students take it literally and believe it. We investigate the effectiveness of different deception techniques through robot teaching. We conduct an experiment in a Japanese Junior High School with 14 students where we investigate the learning and deception effectiveness, and believability using the social robot "Furhat". Moreover, we vary the social agency of the robot by using two different faces, a human (social identity theory) and an anime face (Japanese anime culture). A robot with an anime face achieved a significantly higher learning effectiveness compared to a robot with a human face. However, the human robot face was found to be excellent at deceiving when the paltering deception technique was utilized.

This paper discusses a machine learning approach for classifying student emotions while doing programming exercises. Detection of academic emotions in programming from face features has previously been shown to be a difficult task because people don't tend to display as much expression as compared to more social activities. In our approach, we show that adding log features in addition to face features can improve the performance of classifiers. Furthermore, we show that identifying representative intervals of each emotion type based on human annotations can be used to build models to classify emotion over longer periods of time. We believe that our study can contribute in the development of better intelligent programming tutors that can respond to the affective state of students.

We performed an analysis of the behaviors of novice students while solving programming exercises from data collected in two environments: a controlled laboratory setup and an online system that could be used freely in the wild. We modeled student behavior from raw action sequences using hidden Markov models to capture the sequential information of the problem solving process. We found similar sequential structures between the two environments, with students generally starting from being idle, followed by writing of the code, followed by testing and submission, which may then transition back to an idle state if not successful. While the models were similar, we found evidences of less persistence in solving the problems on students using the online system compared to those in the controlled setup.

We developed a mixed-reality approach, called Crystal Palace, which provides prop-based passive haptic feedback for the 3DUI contest. In this system, we propose an interface of controlling virtual object with physical tools such as scissors, a spray and a screwdriver as game controllers. Our prop-based approach provides natural visual affordance and passive haptics by associating physical tools with various 3D tasks. Using 360 degrees spherical display, players will be able to use common real-world tools in an intuitive way to interact with 3D objects in a virtual escape room.

In ball games, it is important that the players are able to estimate the position of the other players from a bird's-eye view based on the information obtained from their first-person view. We have developed a training system for improving this ability. The user wears a head-mounted display and can simulate ball games in 360° from the first-person view. The system allows the user to rearrange all players, and a ball from the bird's-eye view. The user can then track the other players from the first-person viewpoint and perform actions specific to the ball game such as passing, receiving a ball, and (if a defense player) following offense players.

We developed a system that generates character animations according to several factors, such as the state of the character and the conditions of the open world, using a phase-functioned neural network (PFNN). PFNN is a new neural network architecture that computes weights for every frame using a phase input. When used for character control, it enables the character to adapt its behaviors to various environments and user inputs, producing superior quality animations as compared to other methods. This study examines an approach that adapts character animations to injuries and accounts for the severity of the injuries. We present an injury representation and its implementation into an existing PFNN system. We tested single and concurrent injuries to various body parts with different severity levels while the character walked and ran. The character adapted realistically to various scenarios accurately representing the effects of its injuries, environment and user inputs, while maintaining a fast prediction time.

We developed a game with a purpose (GWAP), which collects structured data corresponding to adjectives to build a visual effect dictionary. In this system, new semantic links can be acquired. Under the guise of a fighting game, the system encourages the user to vote on the commonsense knowledge associated with an object, because our previous research indicated that the rules of showing appropriate visual effect according to the adjective is related to commonsense knowledge of the target object. This system displays visual effects on the target object and the data structure is updated based on user's vote. This structured data underlies a new type of communication support system that continuously improves visual effects that modify adjectives and objects. In this paper, we discuss the structure of the visual effect dictionary through an experiment. Findings show that GWAP effectively improves the relationship between commonsense knowledge and objects, while creating new linkages via deduction.

We discuss an approach in procedurally generating coding exercises for programming practice without using parameterized questions. We applied our approach in the development of a system that can adjust the complexity of the programming exercises generated based on the presence of confusion on the student. A guide for each exercise can also be presented when confusion is detected. We tested our system on university students. 62.68% and 68.57% of the students found the exercises to be fun and helpful in programming practice, respectively. In addition, we also found that adjusting complexity and providing guides based on confusion has a positive impact on the number of problems solved. This study can lay the foundation for the development of programming intelligent tutoring systems that can generate personalized learning content such as controlling the difficulty of exercises based on the proficiency of the student and presenting exercises that are engaging for the learner.

In this study, a communication learning support system was developed to enable people with typical development (TD), who do not have developmental disorders, to learn the most appropriate way of conversing with those who have autism spectrum disorder (ASD). People with ASD tend to face obstacles in interpersonal relationships because they find it difficult to comprehend conversational implicatures. As a solution, we propose a communication learning support system to enable those with TD to learn how to speak to people with ASD in a way they can understand, including how to best convey conversational implicatures. This learning content simulates the experiences of ASD individuals using dialogues based on real-life examples, in order to teach players why people with ASD cannot understand what they mean and how they can paraphrase their message. This learning content is intended to reproduce the experience of having ASD using a semi-humanoid robot and voice dialogue so that those with TD can easily apply learning content when talking with ASD persons. We changed the robot's reaction based on the extent to which TD individuals understand ASD; we determined their level of understanding using a dialogue corpus and adjusted the learning content accordingly. We explain how the dialogue corpus was compiled, and the process of determining TD individuals' comprehension of ASD.

A tablet PC system using an etymological memorization method with animations can enable Japanese children to memorize English words. The system specifically helps with memorization of compound nouns, which are combination of two simple nouns, for animals. For each noun it shows an animation with a narration produced by a voice synthesizer, followed by an animation showing the compound noun, again with narration. The system was experimentally evaluated by comparing the effectiveness of using animation, pictures, and text.