Skip to main content

“…as humans, it’s easy to envision an object with different attributes. But, despite advances in deep neural networks that match or surpass human performance in certain tasks, computers still struggle with the very human skill of ‘imagination.'”

Imagining a new type of chair, car, building, or other object is something that human beings take for granted. But the same sort of creative task is much more difficult for even the most well-trained AI. But with computer science professor Lauren Itti’s new disentanglement approach, AI algorithms will be able to better extrapolate from their training sets to novel objects they’ve never seen before.

Michael Cox

Michael was born and raised in a small island town off the coast of Seattle. He went to USC, but he’s still a Pacfic Northwesterner at heart—he loves hiking, reading, and watching the Seattle sports teams disappoint him. At Viterbi, Michael counsels high school and transfer students in the application process.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.