Group items tagged

Recently, many competitions in the computer vision domain have been won by huge convolutional networks. In the image net competition, the convolutional network approach halves the error from ~30% to ~15%! Key changes that make this happen: weight-sharing to reduce the search space, and training with a massive GPU approach. (See also the work at IDSIA: http://www.idsia.ch/~juergen/vision.html) This should please Francisco :)

...mmmmm... they use 60 million parameters and 650,000 neurons on a task that one can somehow consider easier than (say) predicting a financial crisis ... still they get 15% of errors .... reminds me of a comic we saw once ... cat http://www.sarjis.info/stripit/abstruse-goose/496/the_singularity_is_way_over_there.png

I think the ultimate solution is still to put a human brain in a jar and use it for pattern recognition. Maybe we should get a stagiaire for this..?

Abstract: In the last 3000 yr, one significant and rapid increase in the concentration of 14C in tree rings was observed; it corresponds to a gamma-ray energy input of 7x10^24 erg at Earth within up to one year in AD 774/5 (Miyake et al. 2012).

bio-robotics anybody?

WIRING Poke the left silver wire about one millimeter into the roach's thorax, under a wing just behind its head, and secure it with superglue. Cut each antenna to expose a neuron-lined tube. Insert the middle wire one millimeter into the left tube, and the right wire into the right tube. Superglue both wires into place. CONNECT AND COMMAND Hot-glue the circuit board onto the roach's back and plug it into the head connector. After the roach wakes up, press the remote's left button to urge it right, and the right button to move it left. The cyborg will ignore commands after a few minutes. Peel off the circuit board and clip all wires to ensure a long retirement.

The video says it all... made me laugh for a long time

"The researchers also discovered that the robot is very bad at juggling shoes and Coke bottles"... I wonder if that's the future work directions in their IEEE paper.

For all the fans - here is the directors cut version of this great piece of juggleability-research: http://www.youtube.com/watch?v=5eZY6399hTY

Your TV might be watching you :-) ... If you have one

One more reason why we should be much more open about all these EO data we have .... The two citizen scientists, Kian Jek and Robert Gagliano, are listed as authors on the scientific paper recently published. I love this: the digital nature of these data make it far, far easier to analyze the science than it was in the past, and also easier to get the data out to people. Because of this, we have an explosive growth in these kinds of projects. Planet Hunters is great, but then so is Galaxy Zoo, Moon Mappers, Ice Hunters, and so many others. You can find several of these collected at the CosmoQuest website.

Wiktor is also collecting these science games in the ACT-wiki. You should have a look at: http://sophia.estec.esa.int/actwiki/index.php/On-line_Games_4_Science

Oh, I like this one. One more point for swarm intelligence! :)

We could have come up with an inspiration like this !!! As creative as the roots :-) "These male amphibians use their calls to attract the female, who can recognise where it comes from and then locate the suitor. The problem arises when two males are too close to one another and they use their call at the same time. The females become confused and are unable to determine the location of the call. Therefore, the males have had to learn how to 'desynchronise' their calls or, in other words, not call at the same time in order for a distinction to be made."

Impressive ...

Rover Curiosity finds man-made shiny material on Mars. It seems the rover is littering plastic on the planet!

Their planetary protection officers will have some interesting questions to ask ...

I am certain that this is just the first in a series - highlighting the big dilemma of geo engineering: it's so cheap to do ....

On the same subject ...

Nice.

good luck with this....

In 1996 Sokal hoaxed sociologists with his famous nonsense text on political implications of quantum gravity. Can one play a similar game with "researchers" on Boltzmann brains, multiverses, string landscapes or similar? I doubt, this is just reality satire that can't be topped.

Poor Boltzmann ...

Watch out the Rolling Stones...