論文

Advances in technology have made it possible for people who could not participate in work due to difficulties in social participation to work. It is becoming easier for people who have lost their abilities due to the decline in their abilities as a result of aging or due to some kind of disability to enter work remotely. However, the lack of capacity of these people hinders the implementation of the task. In order to compensate for the lack of human ability and promote remote participation in society, the idea of supplementing a part of remote human ability has been proposed as in Cybernetic Avatar. However, it is not easy to automate the capabilities required at work. In this study, the authors propose a remote-controlled avatar method for people who lack the ability to temporarily borrow the ability of others. The study looked at the models needed for these avatars, and created a prototype to see what problems could arise when working. As a result of the verification, three gestures were observed: symbol gesture, motion gesture, and shape gesture. In addition, as the intention arbitration between the two is performed smoothly, it was observed that the intention arbitration changes from the motion gesture to the symbol gesture.

The automation of complex negotiations is required to coordinate various AIs. In complex negotiations, the goals of each agent may not be shared because some agents may benefit from not disclosing their own information during negotiation. This uncooperative situation requires an agent to have the ability to infer the intention of the others from the communication, and also to persuade the others through communication. Such negotiation in hidden identity is a pioneering area on negotiation research. In the past, the authors have proposed an AI competition in the Werewolf game in order to research negotiation in hidden identity. To eliminate noise caused by the difficulty of handling natural language, it is necessary to create a model of communication between agents in such a competition. In this paper, we analyze the elements necessary for communication between agents in the Werewolf game and design the protocol between agents based on this analysis.

The game agents in cooperative game must be not only intelligent but also acceptable to the player to provide a good game experience. A particularly important factor in implementing an acceptable agent is to make the risk-taking tendency closer to the player. People tend to be attracted to people who has same preference and make similar behavior with themselves. Thus, it is estimated that people are also attracted to cooperative game agents who has same tendency. In this study, we implemented two agents with different risk-taking tendencies. Participants evaluated gameplay with each agent. As a result, Participants with a high-risk taking tendency perceived that agents with a high-risk taking tendency understood the rules of the game more. And vice versa. This suggests that it is possible to improve the player's game experience by making the agent's risk tendency similar to the player's one.

In this paper, we examine the effectiveness of the social distance between a virtual agent and users, and the gaze instruction using a display that can be viewed stereoscopically without using any wearable devices. An actual robot cannot always maintain an appropriate interpersonal distance, through nonverbal gestures owing to its limited range of motion. Because large movements may harm humans, the nonverbal gestures of robots are limited in the real world. In this work, 14 participants were asked how far they wanted to move from a robot posing as a museum guide agent. They were also asked to identify the point at which they felt the agent was gazing. There was a significant distance between the initial position and the position to which the participants moved in the first task under two-dimensional (2D) and three-dimensional (3D) scenarios. The participants moved a significant distance in the first task. In the gaze estimation task, however, the error between the 3D and 2D evaluations was significantly lesser, at a point far from the agent.

Hanabi is a cooperative game for ordering cards through information exchange, and has been studied from various cooperation aspects, such as self-estimation, psychology, and communication theory. Cooperation is achieved in terms of not only increased scores, but also reduced cognitive load for the players. Therefore, while evaluating AI agents playing a cooperative game, evaluation indexes other than scores must be considered. In this study, an agent algorithm was developed that follows the human thought process for guessing the AI’s strategy by utilizing the length of thinking time of the human player and changing the estimation reliability, and the influence of this agent on game scores, cognitive load, and human impressions of the agent was investigated. Thus, thinking time was used as an indicator of cognitive load, and the results showed that it is inversely proportional to the confidence of choice. Furthermore, it was found that the mean thinking time of the human player was shortened when the agent used the thinking time of the player, as compared with the estimation of the conventional agent, and this did not affect human impression. There was no significant difference in the achieved score and success rate of the estimation by changing the estimation reliability according to the thinking time. The above results suggest that the agent developed in this study could reduce the cognitive load of the players without influencing performance.

The Automated Negotiating Agents Competition (ANAC) is a yearly-organized international contest in which participants from all over the world develop intelligent negotiating agents for a variety of negotiation problems. To facilitate the research on agent-based negotiation, the organizers introduce new research challenges every year. ANAC 2019 posed five negotiation challenges: automated negotiation with partial preferences, repeated human-agent negotiation, negotiation in supply-chain management, negotiating in the strategic game of Diplomacy, and in the Werewolf game. This paper introduces the challenges and discusses the main findings and lessons learnt per league.

In this research, the authors considered a method of generating an ecosystem of learning between classrooms using robots that children can edit themselves. A robot that allows children to choose faces and motions, as well as create user content and receive feedback on that content, was developed. This robot system was introduced and tested in an elementary school with 164 students. First, the 5th grades 18 children were taught how to set up the robot, and then they learned how to create usable content. From this, 19 robot contents were created. Next, the robot system was monitored for 2 months, and the teacher-created content was compared with the student-created content. It was found that the content created by the children was more interesting to the students than the teacher created content. Through an interview with the school librarian, the authors found that the classroom has enough learning ecosystem with elderly to younger children.

After analyzing almost all programs pertaining to winning complete information games, game study started to focus on the challenge of programming acceptable game players. A key factor required to create an acceptable player agent is the knowledge of the user’s internal state. People like those who want to know themselves, especially those who want to know their unconscious self-state, also called the “blind self.” This study proposes an approach to analyze an agent’s ability for estimating a user’s blind self as an improvement in human–agent interaction. The authors chose a card game called Hanabi as an evaluation of this type of recursive intelligence. Hanabi is a cooperative and an incomplete information card game in which each player’s cards are unknown to other players. A player decides whether to build a set of cards or discard them while providing hints to the other players. Previous research has demonstrated that the imitation of human behavior that corrects incomplete information increases the score in a Hanabi game between agents. The authors evaluated the agent’s function to modify incomplete information based on the behavior of the players in the game between a human and an agent. The authors experimented with humans and two types of agents. The latter differed according to the need to imitate and modify incomplete information as a cooperator among the experiment’s participants. Afterward, the game results were analyzed, and the impressions of the participants were evaluated. The results demonstrated that while the ability of estimation of the blind self is ineffective for increasing the score, the user developed a good impression of the agent when incomplete information was modified properly, resulting in a similar feeling in the case of a human collaborator.

The authors investigated the effect of gaze tracking by agent for human. Psychological study has shown that presence of others facilitates efforts of working partner. This effect also occurs if the presence of others replaced to robot. Furthermore, previous research shows that the effects on a person being touched by a robot to motivate him or her. However, direct contact such as touch with robot has a risk of disturbing human work. The authors use gaze tracking for social facilitation, because gaze tracking can also facilitate human efforts. The authors implemented social facilitation robot agent by using gaze tracking. The agent projects its face on the surface to show quickly changing expression which is difficult for achieve by ordinally robots. The agent also can express gaze tracking by implemented eye trackers. The authors prepared three conditions with monotonous task for comparing the effect of gaze tracking, first: the agent gazes the aim of task, second: the agent moves its gaze randomly, third: the agent traces the gaze point of participant. The result showed that the participant raise motivation if the agent traces gazed point by human, even though there are no the different on achievement score of task and continuation time.

Surveys show that teleworkers tend to feel isolated due to the lack of informal communication. To solve this problem, we propose a telepresence robot system that triggers informal communication. Drawing on Kendon’s finding of greeting structure prior to interaction, we propose a system that presents two salutations, distance salutation and close salutation. Our experiment with a prototype system suggests that the proposed method is effective in inducing informal communication over the distance.

The authors propose a robotic agent system for a werewolf game to serve as a human-agent interaction method for studying the social behavior of subjects in communication. We developed a robotic agent with a spherical display on the head and created a system that allows people to compete against our robotic agents. We used the Alwolf project framework for our implementation, and applied algorithms extracted from a previous game competition.1.

In the Human Computer Interaction (HCI) field, it has been a long tradition of concern of accessing computer systems by people with visual impairments. It is important to develop high quality user interfaces, accessible, usable, and desirable for these people. In this paper, we first report a preliminary background review about HCI design for people with visual disability. The review of the problems in social interaction the blind people may encounter, are also presented. Further, we narrow down our research scope and focus on gaze and eye contact, which have important social meaning in face-to-face communication. We then identify our research objective that is to design gaze simulation for people with visual impairments. Finally, we report the prototypes of our research project and the progress made so far.