Computers may eventually be able to decipher human emotions, according to researchers at the University of Cambridge.
“We’re building emotionally intelligent computers, ones that can read my mind and know how I feel,” said Peter Robinson, Ph.D., a professor who is leading a team determined to explore the role of emotions in human-computer interaction. His research is highlighted in the film The Emotional Computer.
“Computers are really good at understanding what someone is typing or even saying. But they need to understand not just what I’m saying, but how I’m saying it,” said Robinson.
Human emotions are understood through facial expressions, tone of voice and body language. Even when a person is interacting with a computer, he still puts out these emotional cues, but computers have been unable to decipher them.
The University of Cambridge research team is working closely with Professor Simon Baron-Cohen and the University’s Autism Research Centre team. Researchers there know the difficulties individuals with autism have in understanding emotions, and this knowledge is extremely helpful in the computer-human interaction research.
Since facial expressions are such an important aspect of reading emotions, one system was developed to track a person’s facial features and then analyze facial movements in order to infer emotions from them. It is accurate more than 70 percent of the time, which is about the equivalent of a human observer.
Other emotion-reading computer systems infer emotions by analyzing speech intonation, body posture or gestures.
“Even in something as simple as a car we need to know if the driver is concentrating and confused, so that we can avoid overloading him with distractions from a mobile phone, the radio, or a satellite navigations system,” said Ian Davies, a member of Robinson’s team.
Not only would Robinson like for computers to be able to read emotions, but he also strives for computers to be able to express emotions as well, such as in the form of a cartoon animation or a physical robot. Various team members have been working on creating these forms and giving them believable human expressions.
“The key to scientific research is to avoid preconceptions and to expect surprises. I just recruit the best graduate students from around the world, make sure that they have the resources that they need and then just let them get on with it. They bounce ideas off each other and solve problems together,” Robinson said.
“The team has to combine results from many disciplines, and this is true for many research problems in computer science. We need to understand psychology, signal processing and statistical machine learning as well as systems engineering to tackle these problems. Because the university has experts in all these fields it’s a perfect place to do the research.”
Source: Cambridge University