Science News

IBM And MIT Team Up In A Project To Make Robots Think Like Humans

Photo from Pixabay

The Massachusetts Institute of Technology and IBM Corp. have partnered in an ambitious venture to solve a problem facing computers: teach them to recognize images and sounds in the same way people do, so they can react in useful ways.

The two institutions announced their team-up in what will be a multi-year artificial intelligence program called the IBM-MIT Laboratory for Brain-inspired Multimedia Machine Comprehension. No further details were released on how much the project will require or how many years it will run. But scientists on both sides agree that they face a momentous technical challenge in their goals, the Boston Globe reports.

Guru Banavar, chief science officer for cognitive computing at IBM, says,

We think of people and machines working together to solve problems.

But machines are currently unable to do that and to get there, they must be able to understand human behavior in order to make helpful decisions.

For example, a CCTV camera might one day see that a child has fallen down and recognize that immediate attention or medical help is needed. The camera could then possible transmit the information to family members or law enforcement. The project also aims to let computers learn from other machines, like a robot at a factory learning a new task simply by watching another perform it routinely.

While these traits come naturally to humans, machines don’t have the same abilities. IBM will work closely with MIT’s department of brain and cognitive sciences to discover ways in which this same trait can be transferred onto machines.

James DiCarlo, department head at MIT, says that while there have been “advances in recognizing objects” when it comes to artificial intelligence, there has been little to none in recognizing, much less predicting, human-like actions.

In June, MIT researchers announced that a machine may be able to predict human greetings, able to tell between a handshake or a hug using an algorithm that could anticipate interactions – something that could have useful implications in the future, and similar to what this new project hopes to do.

Click to comment
To Top

Hi - We Would Love To Keep In Touch

If you liked this article then please consider joing our mailing list to receive the latest news, updates and opportunities from our team.

We don't want an impostor using your email address so please look for an email from us and click the link to confirm your email address.