An Intelligence That Can See Its Own Future Created by Scientists

UC Berkeley researchers create a robot that learns by playing and can predict the future of its actions.


Researchers from UC Berkeley have created a technology that allows robots to imagine how their actions will turn out in the future. Using this approach, the robots are able to interact with objects that they’ve never come across previously. 

The way the technology helps the robots predict the future is called visual foresight. It gives the robotic system an ability to anticipate what future actions, like a series of movements, would look like in its camera.

In essence, a robotic imagination is invoked, allowing predictions several seconds into the future. This is not much but lets the robot named Vestri perform tasks without help from humans or much background information about the environment. The visual imagination is improved simply through unsupervised exploration, with the robot playing with objects on a table like a child. That helps the robot build a predictive model of the environment and use to control objects it hasn’t encountered previously. 

Check out this video on how Vestri uses imagination to perform tasks:

Underlying this ability is deep learning based on technology called “dynamic neural advection (DNA)”. Models using DNA predict how pixels in an image will look like in the next frame depending on the robot’s action. Advancements in such video prediction models allow for greater planning capacity for robots as well as performance of complex tasks like changing positions of various objects and sliding toys around obstacles.

Sergey Levine, assistant professor at Berkeley, whose lab developed the technology, thinks robotic imagination can make the machines learn intricate skills.

"In the same way that we can imagine how our actions will move the objects in our environment, this method can enable a robot to visualize how different behaviors will affect the world around it," said Levine. ”This can enable intelligent planning of highly flexible skills in complex real-world situations."

He compared the learning method to how children learn by playing with toys. This research gives robots the same opportunity. 

Chelsea Finn, a doctoral student from Levine's lab, who is also the inventor of the original DNA model, remarked that the significant advancement here is that robots can now learn on their own.

"In that past, robots have learned skills with a human supervisor helping and providing feedback," said Finn. What makes this work exciting is that the robots can learn a range of visual object manipulation skills entirely on their own." 

This tech can find application in self-driving cars that need to predict future events or in robotic assistants. 

Related Articles

Scientists discover what caused the worst mass extinction ever

How a cataclysm worse than what killed the dinosaurs destroyed 90 percent of all life on Earth.

Credit: Ron Miller
Surprising Science

While the demise of the dinosaurs gets more attention as far as mass extinctions go, an even more disastrous event called "the Great Dying” or the “End-Permian Extinction” happened on Earth prior to that. Now scientists discovered how this cataclysm, which took place about 250 million years ago, managed to kill off more than 90 percent of all life on the planet.

Keep reading Show less

Why we're so self-critical of ourselves after meeting someone new

A new study discovers the “liking gap” — the difference between how we view others we’re meeting for the first time, and the way we think they’re seeing us.

New acquaintances probably like you more than you think. (Photo by Simone Joyner/Getty Images)
Surprising Science

We tend to be defensive socially. When we meet new people, we’re often concerned with how we’re coming off. Our anxiety causes us to be so concerned with the impression we’re creating that we fail to notice that the same is true of the other person as well. A new study led by Erica J. Boothby, published on September 5 in Psychological Science, reveals how people tend to like us more in first encounters than we’d ever suspect.

Keep reading Show less

NASA launches ICESat-2 into orbit to track ice changes in Antarctica and Greenland

Using advanced laser technology, scientists at NASA will track global changes in ice with greater accuracy.

Firing three pairs of laser beams 10,000 times per second, the ICESat-2 satellite will measure how long it takes for faint reflections to bounce back from ground and sea ice, allowing scientists to measure the thickness, elevation and extent of global ice
popular

Leaving from Vandenberg Air Force base in California this coming Saturday, at 8:46 a.m. ET, the Ice, Cloud, and Land Elevation Satellite-2 — or, the "ICESat-2" — is perched atop a United Launch Alliance Delta II rocket, and when it assumes its orbit, it will study ice layers at Earth's poles, using its only payload, the Advance Topographic Laser Altimeter System (ATLAS).

Keep reading Show less