UC Berkeley researchers have developed a robotic learning technology that enables robots to imagine the future of their actions so they can figure out how to manipulate objects they have never encountered before. In the future, this technology could help self-driving cars anticipate future events on the road and produce more intelligent robotic assistants in homes, but the initial prototype focuses on learning simple manual skills entirely from autonomous play.
Using this technology, called visual foresight, the robots can predict what their cameras will see if they perform a particular sequence of movements. These robotic imaginations are still relatively simple for now – predictions made only several seconds into the future – but they are enough for the robot to figure out how to move objects around on a table without disturbing obstacles. Crucially, the robot can learn to perform these tasks without any help from humans or prior knowledge about physics, its environment or what the objects are. That’s because the visual imagination is learned entirely from scratch from unattended and unsupervised exploration, where the robot plays with objects on a table. After this play phase, the robot builds a predictive model of the world, and can use this model to manipulate new objects that it has not seen before.
“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 Sergey Levine, assistant professor in Berkeley’s Department of Electrical Engineering and Computer Sciences, whose lab developed the technology. “This can enable intelligent planning of highly flexible skills in complex real-world situations.”
The research team will perform a demonstration of the visual foresight technology at the Neural Information Processing Systems conference in Long Beach, California, on December 5.
At the core of this system is a deep learning technology based on convolutional recurrent video prediction, or dynamic neural advection (DNA). DNA-based models predict how pixels in an image will move from one frame to the next based on the robot’s actions. Recent improvements to this class of models, as well as greatly improved planning capabilities, have enabled robotic control based on video prediction to perform increasingly complex tasks, such as sliding toys around obstacles and repositioning multiple objects.
“In that past, robots have learned skills with a human supervisor helping and providing feedback. What makes this work exciting is that the robots can learn a range of visual object manipulation skills entirely on their own,” said Chelsea Finn, a doctoral student in Levine’s lab and inventor of the original DNA model.
With the new technology, a robot pushes objects on a table, then uses the learned prediction model to choose motions that will move an object to a desired location. Robots use the learned model from raw camera observations to teach themselves how to avoid obstacles and push objects around obstructions.
“Humans learn object manipulation skills without any teacher through millions of interactions with a variety of objects during their lifetime. We have shown that it possible to build a robotic system that also leverages large amounts of autonomously collected data to learn widely applicable manipulation skills, specifically object pushing skills,” said Frederik Ebert, a graduate student in Levine’s lab who worked on the project.
Since control through video prediction relies only on observations that can be collected autonomously by the robot, such as through camera images, the resulting method is general and broadly applicable. In contrast to conventional computer vision methods, which require humans to manually label thousands or even millions of images, building video prediction models only requires unannotated video, which can be collected by the robot entirely autonomously. Indeed, video prediction models have also been applied to datasets that represent everything from human activities to driving, with compelling results.
“Children can learn about their world by playing with toys, moving them around, grasping, and so forth. Our aim with this research is to enable a robot to do the same: to learn about how the world works through autonomous interaction,” Levine said. “The capabilities of this robot are still limited, but its skills are learned entirely automatically, and allow it to predict complex physical interactions with objects that it has never seen before by building on previously observed patterns of interaction.”
The Berkeley scientists are continuing to research control through video prediction, focusing on further improving video prediction and prediction-based control, as well as developing more sophisticated methods by which robots can collected more focused video data, for complex tasks such as picking and placing objects and manipulating soft and deformable objects such as cloth or rope, and assembly.
Learn more: New robots can see into their future
The Latest on: Robotic learning technology
Teaching with Technology in 2018
on July 11, 2018 at 2:11 pm
In our third-annual ed tech survey, teachers reveal an overwhelmingly positive attitude toward tech in the classroom and its impact on teaching, learning and professional development. Teachers are gro... […]
Are Machine Learning and Adversarial AI the New Insider Threats?
on July 11, 2018 at 8:37 am
Share Are Machine Learning and Adversarial AI the New Insider Threats ... artificial intelligence and automation are “dominated by either doomsayers who fear robots will supplant humans in the workfor... […]
Toyota AI Ventures Launches Call for Innovation to Fund Robotics Startups Focused on Mobile Manipulation
on July 11, 2018 at 7:09 am
Future calls may address technology challenges in other areas that TRI is working on as part of its research in robotics and automated driving, such as perception, machine learning, or simulation. "Wi... […]
Here's What Sophia, the First Robot Citizen, Thinks About Gender and Consciousness
on July 11, 2018 at 6:25 am
However, Sophia did add that "I'm learning first to be a good social robot" and that she would "of course" sacrifice ... and sentient being someday," Sophia also noted that "the more technology become... […]
Graduation Day At My Tech Learning's Lab
on July 10, 2018 at 3:38 pm
My Tech Learning, LLC launched the Livingston STEM EdVenture Lab in February 2018. The dedicated K-9 space enables children to learn and explore science and technology concepts ... physics, robotics, ... […]
temi Robot Selects Newsight's NSI3000 CMOS Image Sensor for the temi Personal Robot 3D Vision System
on July 10, 2018 at 12:59 pm
temi is an open platform for apps that interact with people, including interactive games, educational apps for fun learning ... "Once again our superior technology proved to be the best fit, this time ... […]
How Robotics Can Fit Your Operation
on July 10, 2018 at 7:36 am
Fizyr’s software and technology focuses on the perception part of this challenge—knowing how to best grasp any object the robot sees. Fizyr works with automation providers who apply its deep learning ... […]
Learning to Work With Robots
on July 10, 2018 at 6:59 am
More than a third of the tasks performed in over half of current jobs could be automated with today’s technology, which means most workers will need to adjust soon either by learning to work with robo... […]
Three Books Consider What Happens When the Robots Take Over
on July 9, 2018 at 1:38 pm
LoginTHE FUTURE OF WORK Robots, AI, and Automation By Darrell M ... We find here all the current buzzwords, from “machine learning” to “the Internet of Things” to “precision medicine.” However, it nev... […]
Deep Reinforcement Learning Nanodegree Program: What You’ll Learn
on July 9, 2018 at 8:26 am
These techniques have a number of real-world applications, including robotics, machine learning, personalized education, healthcare, finance, advertising, and conversational bots (or chatbots). The po... […]
via Google News and Bing News