Group items matching

in title, tags, annotations or url

Basically neuroscientists discovered time delay neural networks. Guess we were right on track with AI back in the 70s.

TED talks are so random even recurrent neural networks can do it

:D this is so good !!

my favourite sentence: "or to be able to solve the data in an astronaut a spider"

Loosely related to an old ACT project on optical flow (if I remember correctly but even if not still an interesting read I think): "At first glance, the eyes of mammals and those of insects do not seem to have much in common. However, a comparison of the neural circuits for detecting motion shows surprising parallels between flies and mice. Scientists have learned a lot about the visual perception of both animals in recent years."

A neuronal network trained to paint the scene of a given photograph in the style of Kandinsky, van Gogh, or Munch. Their results look quite impressive. Unfortunately the article is in German, but the English paper (with plenty of pictures) is here: http://arxiv.org/pdf/1508.06576v2.pdf Malen wie Kandinsky, wie van Gogh, wie Munch nur auf Basis einer Fotovorlage? Natürlich gibt es begabte Kunstfälscher, die das können. Jetzt aber gelingt es auch Computern, und zwar auf höchst eindrucksvolle Weise. Drei Forscher von der Universität Tübingen haben es geschafft, einem sogenannten künstlichen neuronalen Netzwerk das Malen beizubringen.

Impressive stuff indeed. Paper came out one week ago. Multiple independent implementations have popped out since then: * https://github.com/Lasagne/Recipes/blob/master/examples/styletransfer/Art%20Style%20Transfer.ipynb * https://github.com/jcjohnson/neural-style * https://github.com/kaishengtai/neuralart

they use invasive BCI for accessing to single neural cell on patiences with epilepsy.

why? I think they used patients with epilepsy because they already had an implants with electrodes, for other reasons. Otherwise in normal subjects they could not obtain these results. I wonder who (among normal subject) would like to have electrodes implanted in the brain....thus they use subjects that already have an implant.

well if I got it correctly then the had to cut and fizzle around in these brains ... always scaring, and then they claim that we have in our brains in single neurons abstract images e.g. of marilyn monroe . they speak even of Marilyn Monroe neurons :-) similarily his question "who is in control? are we in control of our neurons or our neurons in control of us? " ....

Good critic to a non neuromorphic approach?

is this really worth a nature paper??

Funny to read this question exactly from you, the all and ever fan of anything linked to bio :-) I have read worse papers in nature and in addition it's just "Nature physics", viz. "Nature garbage." Could be that they don't find enough really good stuff to publish in all their topical clones of Nature.

francesco already posted this below

memory cloning

"It would be very easy not to co-operate, and then it wouldn't work", that's still the important part. I'm sure Dario LOVES this paper. Would be nice to have a coffee with him right now...

well a bit misleading title...

An short article about something called biomimetics. To all I understand, it is not.

nice study - reminds me of what we wanted to do with joga and hibernation ...

Oh yes somebody seems to have found a link between CMS and Biomimetics. Everything is possible now, even Fundamental physics !

Neural networks = biomimetics. That's the conclusion from the TOC. It seems that biomimetics becomes the worse usurper than string theory.

Might be interesting for Giusy...

From the alternative computing series...