News Archives

RSS Feed

  • UNM
  • >Home
  • >News
  • >2011
  • >April
  • >[Colloquium] Robots in the Wild: A General Learning Agent for Natural World Interaction

[Colloquium] Robots in the Wild: A General Learning Agent for Natural World Interaction

April 28, 2011

Watch Colloquium: 

M4V file (676 MB)

  • Date: Thursday, April 28, 2011 
  • Time: 11:00 am — 11:50 am 
  • Place: Mechanical Engineering 218

Brandon Rohrer
Sandia National Labs

Autonomous robots are good at spot welding auto bodies and vacuuming. In most instances where the environment is predictable and the task in known beforehand, they do well. But when the environment is unfamiliar and the task must be learned, they tend to fail, sometimes spectacularly. A brain emulating cognition and control architecture (BECCA) was developed to try to solve this problem. In this talk, I will be describing how BECCA works, the pilot tasks it has been applied to, and the implications of successfully creating a general learning agent.

Bio: Brandon Rohrer: Machines that appear to think and move on their own have fascinated Brandon since he was a child. He has pursued this interest through mechanical engineering degrees at BYU (BS ’97) and MIT (MS ’99, PhD ’02) and through his research in the Intelligent Systems, Robotics, and Cybernetics Group at Sandia National Laboratories in Albuquerque, NM. Current research topics include biomimetic machine learning and automated exploratory robots.