Trouble recognizing the emotional states of people around them is a reality to many children with autism spectrum disorders. To counteract this, many therapists opt for a kid-friendly robot that can demonstrate emotions and is capable of engaging children in the process of emotion imitation. Although this is the best therapeutic strategy, it can only work effectively if the robot is capable of interpreting the child's own behavior. –
Now, researchers at MIT Media Lab have created a personalized machine learning that will work to assist robots in estimating the amount of engagement and interest of each child during emotion imitation therapy interactions. "The long-term goal is not to create robots that will replace human therapists, but to augment them with key information that the therapists can use to personalize the therapy content and also make more engaging and naturalistic interactions between the robots and children with autism," explains Oggi Rudovic, a postdoc at the Media Lab and first author of the study.
Watch Personalized Machine Learning for Robot Perception of Affect and Engagement in Autism Therapy:
"The challenge of creating machine learning and AI [artificial intelligence] that works in autism is particularly vexing because the usual AI methods require a lot of data that are similar for each category that is learned. In autism where heterogeneity reigns, the normal AI approaches fail," says Rosalind Picard, a co-author of the study and a professor at MIT.
During robot-assisted therapy, a therapist will show flashcards exhibiting different emotions to the child. The therapist will then program the robot to express each emotion shown to the child. Through this therapeutic process, they will be able to recognize expressions of fear, sadness, or joy. "Therapists say that engaging the child for even a few seconds can be a big challenge for them, and robots attract the attention of the child," says Rudovic, explaining why robots have been used in this type of therapy. "Also, humans change their expressions in many different ways, but the robots always do it in the same way, and this is less frustrating for the child because the child learns in a very structured way how the expressions will be shown."
The MIT research team used a form of machine learning known as deep learning, a hierarchal-multiple layer of data processing which serves to be useful for autism therapy robots. Deep learning will allow the robots to interpret children's behavior more naturally. "In the case of facial expressions, for instance, what parts of the face are the most important for estimation of engagement?" Rudovic says. "Deep learning allows the robot to directly extract the most important information from that data without the need for humans to manually craft those features." The researchers were also able to probe how deep learning machinery made its estimations, which showed cultural differences between children. "For instance, children from Japan showed more body movements during episodes of high engagement, while in Serbs large body movements were associated with disengagement episodes," explains Rudovic.
Source: Massachusetts Institute of Technology, Science Robotics