Generative adversarial networks

Deep learning algorithms for predicting machine failures in industrial settings can be compressed without compromising their performance, say A*STAR researchers.

A new generative adversarial network-based framework generates artificial brain wave data that improves classification for better brain-computer interface performance.

A generative adversarial network has been used to develop audio classification technologies that require much less training data.

A promising new computational model can recreate clean images out of hazy ones.

A*STAR scientists have developed a generative adversarial network that can accurately classify clothing images and realistically manipulate specific image attributes.