DeepMind’s Neural Network Teaches AI to Reason About the World

By New Scientist

June 15, 2017


A robotic homage to Rodin's "The Thinker."

Google’s DeepMind has developed a neural network that helps computers understand how different objects relate to each other.


Google’s DeepMind has developed a neural network that gives computers the ability to understand how different objects are related to each other.

DeepMind’s neural network specializes in abstract reasoning and can be plugged into other neural networks to give them a relational-reasoning boost.

The researchers say they trained the artificial intelligence system using images depicting three-dimensional shapes of different sizes and colors. The system analyzed pairs of objects in the images and tried to determine the relationship between them.

In addition, the researchers say the system demonstrated its versatility by answering 95% of questions about a set of very short stories related to the pictures.

“There is a lot of work needed to solve richer real-world datasets,” says DeepMind researcher Adam Santoro.

From New Scientist
View Full Article


Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA


No entries found

from Communications of the ACM: Artificial Intelligence


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s