Sign In

Communications of the ACM

ACM TechNews

­CLA Professors Stride Forward in Robotic Technology Research


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
Artificial intelligence systems can help robots learn from their environments.

Statistics and computer science professor Song-Chun Zhu creates robots that learn by collecting data from their surroundings.

Credit: Jesse Wang/Daily Bruin

University of California, Los Angeles (UCLA) computer scientists say recent artificial intelligence (AI) innovations are pushing robotic technology forward.

UCLA professor Song-Chun Zhu produces robots that collect images and other environmental data from digital cameras to learn about their surroundings and feed it into statistical models so they can correctly recognize objects they encounter. The method uses digital image processing and natural-language processing to parse images into basic components to draw conclusions about the object's characteristics. Zhu says it enables AI systems to employ a small set of rules to understand a diverse array of situations, similar to the human learning process.

Armed with such models, Zhu's robots can execute high-level tasks and understand complex notions. However, Zhu says more sophisticated AI intelligence is limited by the accuracy of those models, and a computational model of how people think is needed to get machines to conduct human-like tasks.

Most AI scientists are skeptical that anything approaching such a level of robotic intelligence will emerge in the foreseeable future. Still, Zhu and others think current AI systems may prosper when used for specific jobs that demand little advanced reasoning.

From Daily Bruin
View Full Article

 

Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA


 

No entries found