MIT new algorithm makes robots smart enough to adapt, excel in new environments
The new algorithm helps robots practice skills like sweeping and placing objects, potentially improving their performance at important tasks in houses, hospitals, and factories.
In a breakthrough that seems straight out of science fiction, a team of innovative minds at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), along with The AI Institute, have recently introduced an extraordinary solution set to revolutionize the way robots acclimate and elevate their functionality within new environments.
This pioneering advancement paves the way for robots to seamlessly adapt and improve, promising an exciting future of robotic technology integration into our everyday lives.At last month’s Robotics Science and Systems Conference, researchers presented the “Estimate, Extrapolate, and Situate” (EES) algorithm, enabling robots to learn and improve their skills autonomously.
This innovative approach has the potential to significantly enhance efficiency in various settings, from factories to homes and hospitals.
Getting robots better at work
To improve the performance of tasks such as floor sweeping, EES utilizes a vision system that identifies and monitors the robot’s environment.
The algorithm then estimates how reliably the robot performs an action, such as sweeping, and determines if practicing more is worthwhile.
EES forecasts the robot‘s performance on the overall task after refining the skill and practicing.
After each attempt, the vision system checks to see if the skill was performed correctly. EES could be useful in hospitals, factories, houses, or coffee shops.According to Nishanth Kumar and his colleagues, using only a few practice trials, EES could help that robot improve without human intervention.
“Going into this project, we wondered if this specialization would be possible in a reasonable amount of samples on a real robot,” says Kumar, co-lead author of a paper describing the work, PhD in electrical engineering and computer science, and a CSAIL affiliate.
“Now, we have an algorithm that enables robots to get meaningfully better at specific skills in a reasonable amount of time with tens or hundreds of data points, an upgrade from the thousands or millions of samples that a standard reinforcement learning algorithm requires.”
Promising results
EES’s aptitude for efficient learning was demonstrated when it was utilized in research trials on Boston Dynamics’ Spot quadruped at The AI Institute.
The robot, which had an arm attached to its back, completed manipulation tasks after practicing for several hours. In one demonstration, the robot learned how to securely place a ball and ring on a slanted table in roughly three hours.
In another, the algorithm guided the machine to improve at sweeping toys into a bin within about two hours.
Both results are an upgrade from previous frameworks, which would have likely taken more than 10 hours per task.
“We aimed to have the robot collect its own experience so it can better choose which strategies will work well in its deployment,” says co-lead author Tom Silver SM ’20, PhD ’24, an electrical engineering and computer science (EECS) alumnus and CSAIL affiliate who is now an assistant professor at Princeton University.
“By focusing on what the robot knows, we sought to answer a key question: In the library of skills that the robot has, which is the one that would be most useful to practice right now?”EES could eventually help streamline autonomous practice for robots in new deployment environments, but it has a few limitations for now.
For starters, they used tables that were low to the ground, which made it easier for the robot to see its objects. Kumar and Silver also 3D printed an attachable handle that made the brush easier for Spot to grab.
The robot didn’t detect some items and identified objects in the wrong places, so the researchers counted those errors as failures.