Selfie 2 Waifu takes photos of people and draws an anime-style portrait. The results are outstanding. In this example, a completely normal-looking person is transformed into a ridiculous cartoon character. See the twitter thread below for more examples.

Selfie 2 waifu.

Official Tensorflow implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation (ICLR 2020)https://t.co/3zFxCp669q pic.twitter.com/KdNn80IzKP

— tkasasagi🐻 (@tkasasagi) April 15, 2020

Based on U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation Based on the official pretrain model Basic principle: we will not save your selfie without your permission, all privacy-related actions will be clearly notified. Contact me: creke at creke.net

These Lyrics Do Not Exist differs from similar projects involving human faces, cats, etc., in that the AI-created results are no more uncanny than the real thing. You give it a topic, and select a mood – very sad, sad, neutral, happy, very happy – and a genre – country, metal, rock, pop, rap, EDM – and tell it to get cracking.

Here is the chart-busting hit, "Boing Boing", exactly as generated by the AI:

Verse 1 You came along you showed me My many seasons of different Boing, autumn, winter, boing You touch my boing

Pre-Chorus And oh, the rain it falls on me Soft boing rain on my skin

Chorus So beautiful with you, oh All of boing I look at you, wonder It was the boing of a boing

Verse 2 Take a walk on my heart Far away from all this Take a walk on my mind Fall in the boing of my life

Pre-Chorus So I am autumn, winter and boing And every month brings a new love

Chorus So beautiful with you, oh All of boing I look at you, wonder It was the boing of a boing

Bridge Gotta have a boing in my step, a dawn in my eye Gotta have genuine devotion

Chorus So beautiful with you, oh All of boing I look at you, wonder It was the boing of a boing Read the rest

GANField is the result (nightmarish, grotesque, occasionally comical) of a generative adversarial network trying to find its way through a corpus of Garfield strips.

Watch as an AI struggles to seam together hundreds of Garfield comics. Created by Daniel Hanley. https://twitter.com/calamardh Music is '4 All Hail the Fishmen' from LISA: The Painful

You have no idea how ... unlonely you are, Garfield!

See also: Will Burke's creepy Garfield drawings. Read the rest

You've seen This Cat Does Not Exist, now behold Hopefully This Cat Does Not Exist, whereby the failings of generative adversarial networks become their horrifying strengths.

I'm a big fan of the ones where GANs try to write meme captions and ends up with this ... abyssal syllabary. AI text generation naturally focuses on creating readable phrases, but training it on the appearance of text is where the cold magic springs.

Read the rest

Shardcore (previously) writes, "I took Godley & Creme's seminal 1985 video and sent it through a StyleGAN network." Read the rest

SinGAN is a generative AI that can, among other things, create random variations of an image (or animate it) without knowing what it is an image of. The official implementation is on Github and shows pretty landscapes:

It can also accomplish useful design tasks such as harmonizing color gradients and adding illusory resolution to blurred images.

But I feel that a humorous yet disquieting portrait of the President of the United States of America is in order, courtesy of Jonathan Fly.

SinGAN has never seen a human before, or any other image at all, but it tries its best to animate whatever you give it. It almost succeeds with the hair but then it gets into some genuinely perturbing territory. pic.twitter.com/Wlt28IiaUa

— Jonathan Fly ? (@jonathanfly) November 5, 2019

Read the rest

Michael Friesen generated these abominable pokemon sprites. Be sure to see a similar set hand-drawn by iguanamouth. [via Janelle Shane]

Here's a human illustrating some neural net-generated pokemon names and abilities. Humans earnestly engaging with AI bizarreness is one of my favorite things.https://t.co/bMxUoXSbGB pic.twitter.com/RarbkeJfGX

— Janelle Shane (@JanelleCShane) May 12, 2019

Uncurated samples: https://t.co/CEcXvViZXlPre-trained model: https://t.co/nuXdzwGQtS

— Michael Friesen (@MichaelFriese10) May 13, 2019

Read the rest

Generative Adversarial Networks use a pair of machine-learning models to create things that seem very realistic: one of the models, the "generator," uses its training data to make new things; and the other, the "discerner," checks the generator's output to see if it conforms to the model. Read the rest

Researchers at NYU and U Michigan have published a paper explaining how they used a pair of machine-learning systems to develop a "universal fingerprint" that can fool the lowest-security fingerprint sensors 76% of the time (it is less effective against higher-security sensors). Read the rest