Laughter comes in many forms, from a polite chuckle to a contagious howl of joy. Scientists are now developing an artificial intelligence system that aims to recreate these nuances of humor by laughing the right way at the right time.
The team behind the laughing robot, called Erica, says the system can improve natural conversations between people and AI systems.
Dr. Koji Inoue, of Kyoto University, and lead author of the book Research published in Frontiers in Robotics and AI. “So we decided that one way the bot could empathize with users is by sharing their laughter.”
Inoue and his colleagues set out to teach their AI system the art of laughing during conversation. They collected training data from more than 80 speed dating dialogues between male college students and a robot, initially contacted by four amateur actresses.
Dialogue data was annotated for solo laughter, social laughter (where humor is not included, such as polite or awkward laughter), and playful laughter. This data was then used to train the machine learning system to decide whether to laugh or not, and to choose the appropriate type.
It may seem socially awkward to imitate a small chuckle, but sympathetically to join the hearty laugh. Based on the audio files, the algorithm learned the basic characteristics of social laughter, which tend to be quieter and more playful, with the goal of reflecting them in appropriate situations.
“The biggest challenge we faced in this work was identifying the actual cases of shared laughter, which is not easy because as you know, most laughter is not shared at all,” said Inoue. “We had to accurately categorize the laughs we could use for our analysis and not just assume that any laughter could be answered.”
The team tested Erica’s “sense of humor” by creating four short dialogues to share with someone, and incorporating the new shared laughter algorithm into existing chatbots. These were compared to scenarios in which Erica did not laugh at all or was socially laughing every time she detected laughter.
The clips were run on 130 volunteers who rated the shared laughter algorithm as the most favorable for empathy, naturalness, human similarity, and understanding.
The team said laughter could help create robots with their own distinct personality. “We think they can demonstrate this through their conversational behaviours, such as laughter, eye gaze, gestures and speaking style,” Inoue said, though he added that it could take more than 20 years before there could be an “accidental” chatting with a bot. Like we do with a friend.”
Professor Sandra Wachter, from the Oxford Internet Institute at Oxford University, said: “One of the things I have to keep in mind is that a robot or an algorithm will never be able to understand you. It doesn’t know you, it doesn’t understand you, it doesn’t understand the meaning of laughter.”
“They’re not conscious, but they might be good at making you think they understand what’s going on,” she added.
#Scientists #teach #robot #laugh #time #robots