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

Fabio Rovai

GAN trained on GAP clothes with super-resolution

GAN

Bill T. Jones, Google Creative Lab

Detect poses and voice and add the spoken words live in the choreography

PoseNet, Camera, Voice detection, Microphone

Sofia Crespo

CNN generated blend of corals

CNN

Sterling Crispin

A Facial Recognition model guides a 3D genetic algorithm to create a mask. The masks are 3D-printed with nylon

Facial Recognition, Genetic Algorithm

Christine Sugrue, Google Arts & Culture Lab

Use Object Detection to explore details in artworks

CNN, Object Detection

Anna Ridler

A voiceover via a CNN and RNN for Walt Disney's Nature's Half Acre (1951) trained on text from female protagonists in romance novels

RNN, LSTM, CNN

Zihou Ng

A browser-based exploration of real-time latent space navigation of GAN at higher resolutions with WebGL

GAN

Will Stedden

Translate artist's 3500-word story into a video using CLIP-based Story2Hallucination

Story2Hallucination, CLIP, BigGAN

Golan Levin, Lingdong Huang

A collection of “ambigrammatic” images, generated with a GAN, legible as faces in both upside-down and right-side up orientations

GAN, StyleGAN2, Super-resolution

Albert Barqué-Duran

Use a 3D GAN to generate an Emoji, 3D print the generated Emoji, and paint it during a live performance

GAN, Height map

Maxime Plantady, Evelyne Drouin

Light projections on a sculpture based on facial expressions

Expression Detection

Parag K. Mital

Extract features from 4.7M audio segments and create a tool to play and visualize the audio files

t-SNE

Eirini Malliaraki

Use a pre-trained CNN to expose bias in machine learning-based recruiting processes

CNN, Face markers

Jürgen Schmidhuber, Douglas Eck

Blues Improvisation with an LSTM

RNN, LSTM

Guillaume Slizewicz

A character generator model is trained on a technical database to identify criminals from 1907. It's contrasted with a modern webcam linked to a facial recognition algorithm

Face recognition

Google Brand Studio

Use a camera to find objects and a CNN identifies if it's correct

Camera, CNN, Classification

Refik Anadol

A StyleGAN trained on 2 million architecture photos and interpolated

GAN, ESR-GAN, StyleGAN

Bas Uterwijk

Using Artbreeder to translate Jesus and other historical figures to latent vectors, and adjusting the vector to make them real

GAN, Artbreeder

Gene Kogan

Apply Gradient Smoothing to create an infinite pattern from DeepDream

DeepDream, Gradient Smoothing

Bruce MacDonald

Fooling Facial Detection with Fashion

Facial Detection

Ivan Pravdin

StyleGAN blended model from human poses GAN and FFHQ

StyleGAN, GAN

Prafulla Dhariwal, Heewoo Jun, Christine Payne, Jong Wook Kim, Alec Radford, Ilya Sutskever

Generate music with a combination of VQ-VAE and Sparse Transformers trained on 1.2 million songs

VQ-VAE, Sparse Transformer

Cyril Diagne

Use BASNet to extract an object from the environment and OpenCV SIFT to add it into photoshop

BASNet, OpenCV, SIFT, Camera

Tal Shiri, Joel Simon

Train a DCGAN on Perlin Noise to generate shapes, cut the generated shapes from an Artbreeder image, the cutouts and placement are optimized with feature inversion over a face classifier

DLIB, CNN, Perlin Noise, DCGAN, Artbreeder, StyleGAN, BigGAN, GAN

Submit projects here 💅

☞ ML x ART on Twitter

Copyright © 2021 Emil Wallner

All artworks © of their respective owners