June 30, 2014 – Researchers at University of Washington have turned to social media and crowdsourcing to teach their robots how to do jobs. They presented their findings at this year’s Institute of Electrical and Electronics Engineers conference on robotics in Hong Kong.
The focus was finding a way to get a robot to learn the best way to complete a task, not by repetitive teaching, but by exposing it to online sources showing hundreds if not thousands of people doing similar tasks. The robot observed and translated that into completing assigned tasks. And the robot chose well, picking what would amount to best practices.
Why does this work? Because robots have computer minds, and computer minds like lots of data in which to model solutions for problem solving. So the bigger the crowd sourced the more data the robot can analyze and the better it can produce. This best mimicry approach was called “goal-based imitation” by the university researchers.
The online community came from Amazon Mechanical Turk, a crowdsourcing site.
Related articles across the web
Related Posts
A Climate Change Prehistory Event That Could Repeat Itself
Two Space Stories: One That Matters and One That Does Not
Artificial Intelligence And Its Impact On Future Academic Research
Blood Transfusion Compatibility Research Leads to the Discovery of New Protein Markers
Generative AI Techniques Applied to Robotics is Producing Startling Results
The Present And Future Internet of Things For Sustainable Construction
A New Theory About How Life Originated on Earth and Elsewhere
How AI Will Change Predicting The Weather
