Scientists develop deep learning robots to empower children with autism

NEW YORK: MIT scientists have developed a new type of deep learning network that can help robots evaluate the quality of their interactions with children with autism spectrum disorders using their own data. to each child.

Autism Spectrum Disorder is a condition related to brain development that influences how a person perceives and socializes with others, causing problems in social interaction and communication. The term "spectrum" in autism spectrum disorder refers to the wide range of symptoms and severity.

Armed with a "deep learning" custom, the NAO robot, child-friendly, can easily estimate the commitment and interest of each autistic child, using data specific to that individual , based on a study of 35 children with autism.

The new development can make life easier for them.

"The long-term goal is not to create robots that will replace human therapists, but to supplement them with key information that therapists can use to personalize therapeutic content and also make more natural interactions between robots and children with autism, "said Oggi Rudovic, the first author of the study.

The robot's perception of children's responses was consistent with human expert badessments, with a correlation of 60%, according to scientists.

"The Challenge of Creating Machine Learning and Artificial Intelligence [artificial intelligence] that works in autism is particularly frustrating, because the usual AI methods require a lot of similar data for each category learned.In autism where heterogeneity reigns, the normal approaches to AI fail, "said Rosalind Picard, co-author of the study.

Robot-badisted therapy involves a human therapist showing flash cards of different faces expressing different emotions, such as happiness, sadness, fear and a programmed robot or NAO, which shows the same emotions to the child .

The therapist observes how the child interacts with the robot and gets feedback on how the lesson should be taken further.

NAO is a two-foot robot that looks like an armored superhero or droid, transmitting different emotions by changing the color of his eyes, the movement of his limbs and the tone of his voice.

"Therapists say that the child's engagement for a few seconds can be a big challenge for them, and the robots catch the child's attention," Rudovic said .

"In addition, humans change their expressions in many ways, but robots always do the same, and this is less frustrating for the child because the child learns in a very structured way. how the expressions will be shown, "he said.

In addition to being personalized, the system created by these researchers involves an in-depth learning powered by AI.

Deep learning has been used in automatic speech and object recognition programs, which makes it well suited to a problem such as understanding multiple facial, body and voice characteristics . the commitment of a child.

"In-depth learning allows the robot to extract the most important information directly from these data without humans having to manually produce them," Rudovic said.

For therapy robots, a custom framework was built that could learn from the data collected on each child.