Group items tagged

and here comes my standard question: how can we use this for space? fast detection of natural disasters onboard?

Even on ground. You could for example teach it what nuclear reactors or missiles or other weapons you don't want look like on satellite pictures and automatically scan the world for them (basically replacing intelligence analysts).

In fact, I think this could make a nice ACT project: counting seals from satellite imagery is an actual (and quite recent) thing: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0092613 In this publication they did it manually from a GeoEye 1 b/w image, which sounds quite tedious. Maybe one can train one of those image recognition algorithms to do it automatically. Or maybe it's a bit easier to count larger things, like elephants (also a thing).

In HiPEAC (High Performance, embedded architecture and computation) conference I attended in the beginning of this year there was a big trend of CUDA GPU vs FPGA for hardware accelerated image processing. Most of it orbitting around discussing who was faster and cheaper with people from NVIDIA in one side and people from Xilinx and Intel in the other. I remember of talking with an IBM scientist working on hardware accelerated data processing working together with the Radio telescope institute in Netherlands about the solution where they working on (GPU CUDA). I gathered that NVIDIA GPU suits best in applications that somehow do not rely in hardware, having the advantage of being programmed in a 'easy' way accessible to a scientist. FPGA's are highly reliable components with the advantage of being available in radhard versions, but requiring specific knowledge of physical circuit design and tailored 'harsh' programming languages. I don't know what is the level of rad hardness in NVIDIA's GPUs... Therefore FPGAs are indeed the standard choice for image processing in space missions (a talk with the microelectronics department guys could expand on this), whereas GPUs are currently used in some ground based (radio astronomy or other types of telescopes). I think that on for a specific purpose as the one you mentioned, this FPGA vs GPU should be assessed first before going further.

Wellwellwell! I don't know how I have to complain, since I could not yet read the full article, but I'm sure I will :-).

Discussing with myself :) .... He must have forgotten climate change .... or maybe not?

Well, I have to quote one of the 21st century classics on this one: "I'm sorry mr Musk I can't hear you because of all the Latin chanting. In the meanwhile can you hand me another goat to drain? I'm quite behind on my blood pentagram drawing" - Paul N., AD 2014.

Mr. Musk declined to comment :P

given that they apparently can't decide between two moral options (probably such as humans), should we be more concerned about their stupidity or their intelligence http://arxiv.org/abs/1411.2842

I personally don't even trust people to make "moral" choices in conditions of warfare. History is way too full of examples.

Yes, Black Metal is very good - pure hate! \m/

I always feel like people make too big a deal out of entanglement. In my opinion it is just a combination of a conserved quantity and an initial lack of knowledge. Imagine that I had a machine that always creates one blue and one red ping-pong ball at the same time (|b > and |r > respectively). The machine now puts both balls into identical packages (so I cannot observe them) and one of the packages is sent to Tokio. I did not know which ball was sent to Tokio and which stayed with me - they are in a superposition (|br >+|rb >), meaning that either the blue ball is with me and the red one in Tokio or vice versa - they are entangled. So far no magic has happened. Now I call my friend in Tokio who got the ball: "What color was the ball you received in that package?" He replies: "The ball that I got was blue. Why did you send me ball in the first place?" Now, the fact that he told me makes the superpositon wavefunction collapse (yes, that is what the Copenhagen interpretation would tell us). As a result I know without opening my box that it contains a red ball. But this is really because there is an underlying conservation law and because now I know the other state. I don't see how just looking at the conserved quantity I am in a timeless state outside of the 'universe' - this is just one way of interpreting it. By the way, the wave function for my box with the undetermined ball does not collapse when the other ball is observed by my friend in Tokio. Only when he tells me does the wavefunction collapse - he did not even know that I had a complementary ball. On the other hand if he knew about the way the experiment was conducted then he would have known that I had to have a red ball - the wavefunction collapses as soon as he observed his ball. For him it is determined that my ball must be red. For me however the superposition is intact until he tells me. ;-)

Sorry, Johannes, you just develop a simple hidden-parameters theory and it's experimentally proven that these don't work. Entangeled states are neither the blue nor the red ball they are really bluered (or redblue) till the point the measurement is done.

Hm, to me this looks like a bad joke... The "emergent time" concept used is still the old proposal by Page and Whotters where time emerges from something fundamentally unobservable (the wave function of the Universe). That's as good as claiming that time emerges from God. If I understand correctly, the paper now deals with the situation where a finite system is taken as "Mini-Universe" and the experimentalist in the lab can play "God of the Mini-Universe". This works, of course, but it doesn't really tell us anything about emergent time, does it?

Actually, it has not been proven conclusively that hidden variable theories don' work - although this is the opinion of most physicists these days. But a non-local hidden variable would still be allowed - I don't see why that could not be equivalent to a conserved quantity within the system. As far as the two balls go it is fine to say they are undetermined instead of saying they are in bluered or redblue state - for all intents and purposes it does not affect us (because if it would the wavefunction would have collapsed) so we can't say anything about it in the first place.

Non-local hidden variables may work, but in my opinion they don't add anything to the picture. The (at least to non-physicists) contraintuitive fact that there cannot be a variable that determines ab initio the color of the ball going to Tokio will remain (in your example this may not even be true since the example is too simple...).

The affiliation is Saudi Arabia and Marocco, not countries famous for their contributions to physics... But nonetheless, yes this is possible, to me it even looks very plausible! But you should know that the term "mass" in this context just means a certain parameter in the dynamical equations and only has a loose relation to what we usually call "mass" in the macroscopic world.

ok - admit that I only read the abstract but to me the seems to be a little bit of magic happening ... even if "mass is only a certain parameter in the dynamical equations" ... I assume it still bears some "heavy" consequences in terms of their speed, interactions etc, no? and assuming that you gradually bend such a structure from a 2D to a 1D one ... does it "gain" mass gradually? all very strange to me ...

Amazing how some guys from some other university also did pretty much the same thing (although they didn't use the bidirectional stuff) and published it just last month. Just goes to show you can dump pretty much anything into an RNN and train it for long enough and it'll produce magic. http://arxiv.org/pdf/1410.1090v1.pdf

meanwhile, at Google... http://www.technologyreview.com/view/532886/how-google-translates-pictures-into-words-using-vector-space-mathematics/

Seems like quite the trend. And the fact that google still tries to use LSTMs is even more surprising.