Celebrity power couples don’t often last too long, but the magic of their romance sometimes has us wishing they could have had a fairy tale ending. Now, with the help of state-of-the-art artificial intelligence, we can imagine a more idyllic future where the power couple stayed together and had children instead of publicly catastrophic splits. Here we have used StyleGAN, an AI computer vision algorithm, to prompt an AI to “imagine” what celebs’ adult children would have looked like all grown up.
Divorced and Never Had Kids
Jennifer Aniston and Brad Pitt
Miley Cyrus and Liam Hemsworth
Nick Cannon and Mariah Carey
Jessica Simpson and Nick Lachey
Jennifer Aniston and Justin Theroux
Miranda Lambert and Blake Shelton
Russell Brand and Katy Perry
Broke Up and Never Had Kids
Britney Spears and Justin Timberlake
Ben Affleck and J-Lo
Rihanna and Drake
Kylie Jenner and Tyga
Kristen Stewart and Robert Pattinson
Demi Lovato and Joe Jonas
Taylor Swift and Taylor Lautner
Zac Efron and Vanessa Hudgens
Justin Beiber and Selena Gomez
Bill Clinton and Monica Lewinsky
Leonardo DiCaprio and Kate Winslet
The Office’s Pam and Jim
The Notebook – Rachel McAdams and Ryan Gosling
Bradley Cooper and Lady Gaga
The “children” were imagined by an implementation of StyleGAN, a state-of-the-art AI that generates new images using Generative Adversarial Networks that compete to create increasingly realistic-appearing images that remain true to life.
The AI is made up of a generator network and a discriminator network. The generator network creates images, and the discriminator decides if that image is “real” (similar enough to real images it was trained on) or “fake.” This process continues in a loop, progressively creating more and more realistic “imagined” images.
Note: We did not develop the phenomenal algorithm used to generate these faces. The StyleGAN algorithm used to produce these images was developed by Tero Karras, Samuli Laine, and Timo Aila at NVIDIA, based on earlier work by Ian Goodfellow and colleagues on Generative Adversarial Networks (GANs).