Humans not only prefer to speak precisely and to convey information, but also enjoy conversation for its own sake. This type of conversation plays an important role in establishing bonds of solidarity among participants. In this study, a model is developed that can portray various types of conversational atmospheres based on the concept of phatic expression, and reproduce conversations and atmospheres using multiple animated agents. A simulation system is also developed to reproduce conversational atmospheres by controlling animated agents' verbal and nonverbal behaviors. Based on the proposed model, a simulation can generate informative or phatic atmosphere.
雰囲気の研究，多人数エージェントの研究
仲の良い雰囲気を示す抽象表現のデザイン, The Inference of Friendly Communicative Atmosphere Created by Geometric Shapes
Many previous studies on inference of social behaviors using geometric shapes have explored causality, animacy, intention, and desire inferred from the movements of such shapes;
however, inference of communicative atmosphere in terms of friendliness/antagonism using geometric shapes has not yet been studied well.
This study investigated how a friendly/antagonistic communicative atmosphere was inferred from the movement of two egg shapes.
We developed animations for these shapes involving forward/backward/parallel tilts with coincidence/incoincidence of synchronous movement.
Results showed significant differences in the inference of friendly/antagonistic atmosphere between coincident and incoincident synchronous movement.
In addition, the inference of a friendly atmosphere was affected by the combination of forward tilt with incoincident movement, which may be interpreted as interaction between the shapes, such as responding or providing feedback.
This suggests that individuals may infer a friendly/antagonistic communicative atmosphere from both coincident movement and incoincident movement interpreted as an interaction.
会話における雰囲気モデルの提案とエージェントによる再現, Conversational Atmosphere Model and Reproduction by Animated Agents
Humans not only prefer to speak precisely and to convey information, but also enjoy conversation for its own sake.
This type of conversation plays an important role in establishing bonds of solidarity among participants.
In this study, a model is developed that can portray various types of conversational atmospheres based on the concept of phatic expression, and reproduce conversations and atmospheres using multiple animated agents.
A simulation system is also developed to reproduce conversational atmospheres by controlling animated agents' verbal and nonverbal behaviors.
Based on the proposed model, a simulation can generate informative or phatic atmosphere.
多人数キャラクタを用いた社会的場面におけるソーシャルスキル訓練システムの提案, Social Skills Training System Simulating Highly Social Scenes Portrayed by Multiple Animated Characters
In recent years, the development of information technology and the decrease in opportunities for direct communication among young people have led to weakened "social skills", namely, their ability to verbally communicate with others, interpersonal skills, and stress responses.
Systems that simulate social-skills training using an animated character have been developed.
However, the existing systems only assume scenes involving a one-to-one conversation; conflicts or pressure situations like highly social scenes are not considered.
Therefore, we propose a training system that reproduces various kinds of communication scenes using multiple animated characters.
Through role-playing and feedback, the user can experience various interactions among people and can acquire social skills.
デフォルメキャラクタでわかる飲食店の雰囲気,Restaurant-Guide Site Using Atmosphere Expressed by Abstract Characters
In recent years, as use of the Internet has spread and the use of Web services has increased, websites of restaurant guides for gourmet dining have been developed.
However, no guide has been developed that features a restaurant atmosphere, even though people often choose a restaurant in anticipation that the atmosphere is conducive to intimacy and romance.
Therefore, we propose a restaurant-guide site that shows images of restaurant ambiance using the body posture of abstract characters.
Our findings will be useful in investigating perceptions of the atmosphere created by characters.
天使と悪魔とあなた ―ユーザの意思決定を促進させるキャラクタの提案と評価―, Do You Follow Angel Agent's Advice? A Pilot Study for Decision Making Support by Multiple Animated Agents
With the increased use of information technologies,the importance of information services that assist us in one's daily decision-making has also increased.
However, this is not enough to help in all our daily choices.
When conditions are conflicting or complex, it is difficult to make appropriate decisions.
In the case of conflict situation of decision-making, several films, television or manga shows often depict internal conflict by showing an angel and a demon characters that offer competing recommendations.
The expression of two different characters may be used in the decision-making tool by using two types of animated agents that can induce suitable decisionmaking.
Therefore, we proposed the decision support animated agents of angel and demon. In this study, we developed animated agents of both the angel and the demon that expressed their roles and promoted suitable decision making.
The angel and demon agents responded against the user's utterance.
The user may understand both for and against aspects in a conflict situation and then the make better decisions.
We conducted two experiments: one using utility function without animated agent and the second one using only animated agent.
The experimental results showed that when the angel and demon agents were used, the participant selected good choice.
This is a useful way to use multiple characters with different characteristics in conversation.
擬人化エージェント／擬人化キャラクタの研究
動物アシスタントとの複数回インタラクション後の利用継続可能性比較, Do you forgive past mistakes of virtual assistant? A study on changing impressions of virtual character when using its assistance multiple times
We investigated the gain-loss effect of virtual/personal assistant character, which provides intelligent assistance to humans, and focused on not only the first impression of using the assistance but also changing impression about the character when used multiple times for assistance.
The experiment used a fictive retrieval system (searching onomatopoeia), and the virtual assistant character looked up for suitable words for the user (success), or failed to find the words (mistake).
There were three sessions, differing by the task of character's mistake; two tasks were successes and one was a failure in each session.
The results showed that the group of people who had low expectation from its first appearance, formed negative impressions after the final mistake, significantly.
Consequently, final mistake influenced the formation of negative impression more than other mistakes, thus showing that the final mistake in multiple times of assistance was associated with loss effect.
