Start End Activity
12:00 13:30 Lunch and Lab tour (please see HAI2019 workshop page)
NGHAI2019 Workshop starts
13:30 13:40 Opening, 10min
by students at Kyoto Institute of Technology, 40min
14:20 14:30 Break, 10min
14:30 15:15 Oral Session, 45min
15:15 16:15 Keynote Speech, 60min
16:15 16:25 Break, 10min
16:25 17:15 Group Discussion 1, 50min
17:15 17:25 Break, 10min
17:25 18:15 Group Discussion 2, 50min
18:15 18:30 Closing, 15min
NGHAI2019 Workshop ends
18:30 20:00 Get together (please see HAI2019 workshop page)


Students at Kyoto Institute of Technology demonstrate their works created in the class of INTERACTION DESIGN PROJECT.
Demonstration list
"Fro Pro: English L&R Pronunciation Practice Robot"
*Haruka Goi, *Ryosuke Omochi, Milena Kaneko, Emma Huffman, Hadrien Reynaud
"Little Us: A Lamp That Reflects the Conversation State"
*Naruhei Miyao, *Kazuki Oi, Gaishi Kudo, Juri Nabesawa, Kazuki Nishisaka, Shiori Fujisaki
"Brainder: Make Your Own Time with Your Own Blend Coffee"
*Mitsuru Tsunekawa, *Shohei Koda, Hiroto Ichihara, Kosuke Shiromoto, Reika Kumazawa, Takuji Yoshida
"Facilitating Conversations by Vibration Feedback Based on Emotion Recognition from Voice Quality"
*Chatani Sota, Tanaka Kyoka, Huruta Kio, Mistukuni Kazuhiro, Doi Kentaro, Nakano Kanta

Oral Session

10 minutes presentation and 5 minutes Q&A per presenter
Start Authors Title & Abstract
14:30 Ryosuke Takata,
Yugo Takeuchi
(Shizuoka University)
Design of Cooperative Interaction between Humans and Agents through Reinforcement Learning
Currently, it is difficult for humans and AI agents to cooperate. In this paper, we propose the design of cooperative interaction between humans and agents. As an experiment, two agents learned to lift a heavy box in a virtual environment simultaneously. As a result, one of the agents was able acquire the behavior of following the other agent automatically. We need to verify whether cooperation between humans and agents can be established through ongoing investigations.
14:45 Jaeryoung Lee
(Chubu University)
Activity, habits, and personality traits in a socially assistive robot environment
As facing the super-aged society, socially assistive robots for elderly care has been getting attentions for the last years. This work, a part of EU & MIC project Culture Aware Robots and Environmental Sensor Systems for Elderly Support (CARESSES), proposes a system understanding the daily activities, habits, and personality of humans for a personalized Human Agent Interaction environment. A low-cost activity recognition system is developed and its mainly collecting a sequence and a duration of human’s activities. The information provided by the activity recognition module processes for estimating the habits of individuals. The proposed system along with its habit representation uses to assess the aspects of individuals’ personality traits by tracing the association between the habits.
15:00 Kazuki Oi,
Mariko Ogishi,
Hiroyuki Morisaki,
Kotaro Yamaguchi,
Kazuaki Tanaka,
Natsuki Oka,
Yukiko Nishizaki
(Kyoto Institute of Technology)
Dose individual differences in acceptance of others affect response speed in synthetic speech sound?
The previous study (Kubo and Nishizaki, 2018) reported that participants who tend to accept others reacted more slowly to an instruction from an autonomous vehicle by synthetic speech sound than those who tend not to accept others. The previous study considered that participants who have high acceptability of others might carefully listen to the instruction and did not react until the end of the instruction. This study aimed to verify the hypothesis. The main differences between the current study and the previous study were as follows: 1) While participants were asked to solve a calculation task during automatic driving in the previous study, there was no such subtask in the current study; 2) The instruction given in the previous study was to switch to manual operation, but the instruction in the current study was to simply press a key. As a result of verification, no correlation was found between the acceptance of others and the reaction time, so it is thought that the result of the previous research came from other factors such as the existence of a subtask. Participants who tend to accept others might be delayed in responding to the main task because they worked hard on the given subtask.

Keynote Speaker

Prof. Michita Imai, Keio University, Japan
Keynote title: Difficulty of HAI research
    • Imai Laboratory, Keio University
Michita Imai is a professor at Keio University, Japan. His laboratory focuses on the mechanisms of interactive AI which can communicate with humans. His current studies investigate a multi-modal speech dialogue system and a recognition system for the situations of route guidance, advertisement, caretaker, education by using CG agents, robots, and IoT systems.

Moreover, NGHAI workshop requests him to give a talk on the key to apply research fund for young researchers, e.g. how to apply for research funding and how to write a great research application. We are sure the keynote is beneficial for the young researchers and the group discussion after the keynote.

Group discussion

Theme: Construction of new academic field in human-agent interaction researches

In the group discussion, participants brainstorm and discuss issues around HAI and aim to build the new research field of HAI. After the workshop, we are going to apply the discussed idea for research grants with workshop participants.