Submit a Project

ML x ART

A showcase with creative machine learning experiments

👋 I'm Emil, a resident at the Google Arts & Culture Lab. I've curated this collection.

Year
Artist
Technology
Theme
Medium
Default order

Yotam Mann, Google Creative Lab

A tool to mix the latent space of two instruments

VAE, Magenta

Anna Ridler

A GAN trained on handmade ink drawings based on the film Fall of the House of Usher (1929) as well as the GAN-generated images themselves

GAN

Alexander Mordvintsev

Reverse Super-Resolution by turning high-res images into low-res

Super-Resolution

Mario Klingemann

Deleting weights in GANs to create glitches

GAN, Weights

Gene Kogan

Apply Gradient Smoothing to create an infinite pattern from DeepDream

DeepDream, Gradient Smoothing

Gene Kogan

Continuously zoom in a image and apply DeepDream and use Dense Captioning to label objects in the video

DenseCap, DeepDream

Refik Anadol

A StyleGAN trained on images from Mars MRO and interpolated frame-by-frame to create a 3D effect

GAN, ESR-GAN, StyleGAN

Stephanie Dinkins

An RNN trained on oral histories from three generations of women from a single-family to highlight concerns of people who are underrepresented in the tech sector, displayed with an interactive voice assistant

RNN, LSTM, Microphone

François Quévillon

A tool for creating an endless car video. It's made by modifying parameters in the sampling system that's created with image features and other technologies

CNN

Cyril Diagne, Google Arts & Culture Lab

Use UMAP to cluster the color of outfits in several thousand catwalks

Background Removal, UMAP

Jake Elwes

Use an image captioning model that translates CNN features to RNN features, and another GAN that translates RNN features to images in a loop

RNN, Image Captioning, CNN, GAN

Mayukh

Optimize the input image to maximize activations of a certain channel inspried by Olah, et al

Gradient Ascent

Anne Spalter

An image with circles Style Transferred with pictures of three airplanes using Nightcafe, and the spaceman and asteroid are animated with Plotaverse

Style Transfer

Gene Kogan

Video texture synthesis using contentless Style Transfer

Style Transfer, Contentless

Shardcore

First Order Motion applied to Francis Bacon's Figure With Meat using a video of Boris Johnson

First Order Motion

Sonam Damani, Nitya Raviprakash, Meghana Joshi, Puneet Agrawal

Generate GAN images and weave the generated images

GAN

Jack Qiao

Generate font combinations by matching the CNN features of fonts

NNS, CNN

Helena Sarin

SNGAN trained on doodles and upscaled and colorized with CycleGAN

GAN, CycleGAN, SNGAN

Ben Snell

A 3D-GAN trained on sculptures inspired by the influence of nature and nurture

GAN, 3D

Tatiana Zobnina, Alonso VĂĄzquez

Using DDSP to translate a Beluga whale voice into a violin

Spleeter, DDSP, BigGAN

Robbie Barrat, Acne

Train a GAN on Acne clothing, generate images with the GAN, and create physical products according to the GAN images

GAN

Alsino Skowronnek

A Conditional GAN trained on a data set of ~1,000 annotated street tag images from Germany. The model can generate new graffiti tags based on any textual input from the audience

Pix2Pix

Mike Tyka

A GAN trained on faces from Flickr

GAN

Aven Le Zhou

A pix2pix model that translates drawings to Chinese “Shanshui” artworks

pix2pix

Submit projects here 💅

☞ ML x ART on Twitter

Copyright © 2021 Emil Wallner

All artworks © of their respective owners