Cognitive Interaction Design Workshop

Workshop of Cognitive Interaction Design discusses interaction mechanism between people and an artificial artifact. People assume that an unknown artifact has a mental state if it behaves spontaneously and responds to people based on a regulated rule. This workshop has four speakers who are investigating the interaction between people and an artifact or between people in terms of a cognitive mental state. We discuss the possibility of a design theory related to a mechanism of reading people's mental state and expressing the artifact's mental state through the presentations.

Date: Oct. 24, 2015
Time: 9:30am-12:00pm
Place: Room 104, 1F IT2 Building (College of IT Engineering)

1. HAI Design Deriving Transitions of User Internal States
Tetsuya Matsui*、Seiji Yamada**

*Project Researcher, National Institute of Informatics

**Professor,  National Institute of Informatics, SOKENDAI, Tokyo

     Institute of Technology


Our goal is proposing a model of user trust transition in interacting with product recommendation agents in an on-line shopping. In many situation that we interacting with agents, for example in on-line retailing, trust for agents is an important factor. We suggest two factor of trust, user emotion and agent intelligence. In our model user's positive emotion and agent's high intelligence both raise agent's trustworthiness.  Agent can infect positive emotion to user. High intelligence can be shown trough speech.  We presuppose model of user trust transition that can transit by changing agent's emotion and intelligence.  We inspected this model in a participants experiment.

2. Autonomous Wheelchair that Reproduces Degree Expressions for Personal Adaptation

Komei Hasegawa, Michita Imai

Project Researcher, Keio University


We developed DECoReS (Degree Expressional Command Reproducing System) that allows an electric wheelchair to adapt a traveling style to user preferences. When users control a wheelchair through voice commands, they can sometimes give such orders as "go straight speedily" and "curve to the right widely." We define these optional words such as "speedily" and "widely" as "degree expression." Because degree expressions are ambiguous, traveling styles described with such expressions are altered depending on the users and environments. DECoReS realizes the travels suited per user by learning degree expressional commands and traveling data from the users. DECoReS also reproduces travels suited for a current environment by exacting the data with a similar map. Our experiments show that DECoReS can reproduce different travels depending on degree expressional commands, users, and environments.

3. A Computational Model of Children's Imitation Behavior During the Eurythmics,

Hirofumi Hayakawa*, Misa Yoshizaki*, Natsuki Oka*, Takayuki Nagai**, Takeshi Omori***

* Kyoto Institute of Technology

** The University of Electro-Communicaitons

***Tamagawa University

4. Estimation of customer's preference from verbal and on-verbal cues in face-to-face selling situation

Hidehito Honda*, Ryosuke Hisamatsu*, Yoshimasa Ohmoto**, Kazuhiro Ueda*,

*The University of Tokyo,

**Kyoto University


In the present study, we examined human-human interaction. In particular, we examined the interaction between shop clerk and customer in face-to-face selling situation on travel plan. In this situation, shop clerk estimates customer's preference about travel plan and proposes the plan the customer will find attractive. Through experimental study, we examined how shop clerk estimated customer's preference about travel plan base on verbal cues (e.g., what customer said) and non-verbal cues (e.g., gaze, hand gesture, body position), and discussed how these cues affected the estimations of customer's preference.

5. Discussion with speakers and attendees


ACM the 3rd International conference Human-Agent Interaction HAI2015

Cognitive Interaction Design Research Project


Michita Imai Keio University (michita at