With Aloimonos, Fermuller, and Yang, Robots can learn to cook

Because of Professor Yiannis Aloimonos’ (CS, CVL, UMIACS) , Dr. Cornelia Fermuller’s (UMIACS), and PhD candidate Yezhou Yang's research, robots can learn to cook in a way similar to the rest of us.  By watching instructional videos on YouTube, the robot in their project is able to determine what series of steps is necessary to complete a task—in this case, how to manipulate objects associated with cooking and learn to cook.

Their project is being done in conjunction with the National Information Communications Technology Research Centre of Excellence in Australia (NICTA). It is also the subject of a UMIACS article that details their ground-breaking work which they will be presenting at the Association for the Advancement of Artificial Intelligence Conference in Austin, Texas on January 29, 2015.

The work of Professor Aloimonos, Dr. Fermuller, and Yang embodies many of the terrific advances in the fields of Computer Vision, Neural Networks, Automatic Speech Processing as well as Natural Language Processing. Many  significant contribultions are continually made to these area of research at the University of Maryland in The Department of Computer Science, in UMIACS as well in The Center for Automation Research (and in this case, more specifically, The Computer Vision Laboratory).   Aloimonos’ research focuses upon real-time systems that are able to perceive—moving from language to action. Fermuller's research centers on the processes involved when machines obtain information and interpret images from video sequences and from stereo images.

Aloimonos will be teaching graduate students an advanced topics course in this very subject area in Spring of 2015: CMSC828Y Advanced Topics in Information Processing; Vision for Cognitive Robots: Recognition, Navigation and Manipulation.

Please read the full UMIACS article about Aloimonos', Fermuller's, and Yang's research on the CMNS Website.

This work has gained a great deal of traction in the press. Please read and article from Engadget and USA Today.

The Department welcomes comments, suggestions and corrections.  Send email to editor [-at-] cs [dot] umd [dot] edu.