A research team at the University of Southern California has developed an artificial intelligence system with an envisioning capability of never-before-seen objects.
"We were inspired by human visual generalization capabilities to try to simulate human imagination in machines," says Ph.D. student Yunhao Ge, lead author of "Zero-Shot Synthesis with Group-Supervised Learning," presented at the virtual 2021 International Conference on Learning Representations.
"Humans can separate their learned knowledge by attributes and then recombine them to imagine a new object. Our paper attempts to simulate this process using neural networks," Ge says.
The disentanglement approach "truly unleashes a new sense of imagination in AI systems," says co-author and computer science Professor Laurent Itti.
From University of Southern California
View Full Article