neohee

Exploring colour data with t-SNE, #ThreeJS and canvas-sketch.

Nice blog post titled "The Quiet Semi-Supervised Revolution" by Vincent Vanhoucke. It discusses two related works by the Google Brain team: Unsupervised Data Augmentation and MixMatch. towardsdatascience.com/the-quiet-semi…

Face to pix2pix facade

The Substitute - @alexandradaisy's fascinating new artwork - draws on our Grid Cells work published in @nature last year to create a visualisation of the extinct male northern white rhino.

transition between worlds

Any other idea would be possible!! (Object) <-> (Design) It would be a new form of interactive design without using a mouse!

This part of @_ScottEaton_'s talk where adding some shadow lines transforms the output of a pix2pix model is astonishing! The whole video here: youtu.be/TN7Ydx9ygPo

#AI isn't just a technical field anymore. We need educators, policymakers, journalists and the rest of society to contribute their perspectives, said @drfeifei, co-director of Stanford HAI. stanford.io/2udpQpO

Some ways of combining information in two branch of a net A & B: 1) A+B, 2) A*B, 3) concat [A,B], 4) LSTM-style tanh(A) * sigmoid(B), 5) hypernetwork-style convolve A with weights w = f(B), 6) hypernetwork-style batch-norm A with \gamma, \beta = f(B), 7) A soft attend to B, ... ?

@quasimondo I also think that its also a sign that Duchamp land residents feel threat to their artwork. They might want to keep their job by denying computer works

@quasimondo In my case, I tried to find images that can be viewed as 'beautiful' , 'peaceful' ... through latent space. But it was quite hard to find them except 'creepy'. I believe if images can convey various feelings, that issue can be changed..

If you really want to annoy me just label me a "computer programmer" when writing about my art. I didn't even know that there are still computer programmers around, I thought everyone is a coder or a developer these days.

An OpenAI employee printed out this AI-written sample and posted it by the recycling bin: #sample8" target="_blank" rel="nofollow noopener">blog.openai.com/better-languag…

We've trained an unsupervised language model that can generate coherent paragraphs and perform rudimentary reading comprehension, machine translation, question answering, and summarization — all without task-specific training: blog.openai.com/better-languag…

Pretty cool how StyleGAN can generate entire fonts in one shot.

So after training #stylegan 7 days on our photobooth images here is the mandatory interpolation video. Looks like it handles the transitions from singleface to multiface sequences a bit better the pro-GAN ... Still have to play with stylemixing now!

@refikanadol Wonderful! how did it make a 3d shape image?

Dear friends, we are deeply exploring photographic memories with machine intelligence. We are constantly training a GAN algorithm with hand picked ~2 millilons architecture photography datasets to create new machine hallucination data sculptures.

A way that GAN should go #thingsihavedrawn Instagram.com/thingsihavedra…

@quasimondo At the same time, I was sad that even if there's still the role of an artist, it is not like what it used to be. Or maybe it would be the job for the artist to expand their role through technology. By the way, thank you for your sincere reply :-)

@_neohee Just like everyone can buy a camera and can take pictures of the world. There is still an infinite amount of possible images of which only a few are interesting or artistic and the choice of what you decide to frame makes the difference between banality and art.

#BigGAN seems like diminishes the role of the artist tremendously. This makes the art as a maid of the technology. GAN researcher would take of the role of the artist. What should the artist do?

@quasimondo I was thinking about your comment for a couple of days and I think you are right! The output images that you choose from the network is something new, never been shown before for centuries.