- Researchers are attempting to teach robots to laugh for improving human robo interaction.
- In the researchers’ shared-laughter model, a human laugh first, and the AI system responds with empathetic laughter.
- The researchers began by annotating speed-dating dialogues to collect training data.
University of Kyoto researchers are attempting to teach a robot to laugh. The researchers created a method for developing a “sense of humor” for the Japanese android Erica, which is described in a research paper published in Frontiers in Robotics and AI.
Robots cannot detect or even laugh at a joke, so researchers want to recreate the human nuances of humor for an AI system to improve how natural conversations can occur between robots and people.
“We think that one of the important functions of conversational AI is empathy. The conversation is, of course, multimodal, not just responding correctly. So we decided that one way a robot can empathize with users is to share their laughter, which you cannot do with a text-based chatbot.” said lead author Koji Inoue, in a press statement. Inoue is an assistant professor at Kyoto University in the Department of Intelligence Science and Technology within the Graduate School of Informatics.
In the researchers’ shared-laughter model, a human laugh first, and the AI system responds with empathetic laughter. To accomplish this, the researchers needed to design three subsystems: one to detect laughter, another to decide whether to laugh and a third to select the appropriate type of laughter.
The researchers began by annotating speed-dating dialogues to collect training data. Speed dating is a matchmaking process in which people have brief conversations with a large number of people to see if they have any mutual interests. The researchers used data from a speed dating session involving Kyoto University students and the robot Erica, which was controlled by various amateur actresses.
“Our biggest challenge in this work was identifying the actual cases of shared laughter, which isn’t easy because as you know, most laughter is not shared at all. We had to carefully categorize exactly which laughs we could use for our analysis and not just assume that any laugh can be responded to,” added Inoue. Another important thing was to ensure that the AI responds with the right type of laughter. For example, laughing out loud could make things awkward in a situation that only warrants a polite chuckle.
After training the robot on this data, it was time to put Erica’s newly acquired sense of humor to the test. They started by recording four short two-to-three-minute dialogues between a person and Erica, who was now equipped with the shared-laughter system. They also added two sets of dialogues: one in which Erica didn’t laugh at all, and one in which she emitted a “social laugh” whenever she detected laughter.
The clips of these dialogues were then played to 130 volunteers who rated them based on empathy, naturalness, human likeness, and understanding. The shared-laughter system performed better than the other two.
