is a machine learning educational tool that explores the latent space of GAN. 

︎ Visit site

Collaboration with Avner Peled.
Comissioned by Ars Electronica Future Thinking School.

The GAN Explorer is an interactive web-based tool that lets you explore the latent space of StyleGAN V1’s results, using the dataset of Flickr-Faces-HQ Dataset. Same dataset was used at 'This Person Does Not Exist' project. The latent space explorer tool invites you to encode your own images to the latent space, and create chains of key-frame animations between different points on GAN's latent space in a collaborative way.

The tool was developed and designed for Ars Electronica Future Thinking School.  co-funded by the European Commission's DG CONNECT, in the framework of the Horizon 2020 program of the European Union under the S+T+ARTS program's Regional STARTS Centers, and with the support of the National Film Board of Canada.
This project is part of the development of the Marrow project with the support of the National Film Board of Canada, ATLAS V and the CNC.

︎ Medium post

Featured:  Ars Electronica festival 2021. 

︎   ︎