Group items tagged GPU - Advanced Concepts Team

Picture This: NVIDIA GPUs Sort Through Tens of Millions of Flickr Photos - 2 views

blogs.nvidia.com/...flickr

AI

shared by Alexander Wittig on 21 Jul 15 - No Cached

LeopoldS liked it

Alexander Wittig on 21 Jul 15

Strange and exotic cityscapes. Desolate wilderness areas. Dogs that look like wookies. Flickr, one of the world's largest photo sharing services, sees it all. And, now, Flickr's image recognition technology can categorize more than 11 billion photos like these. And it does it automatically. It's called "Magic View." Magical deep learning! Buzzword attack!

<div class="cArrow"> </div><div class="cContentInner">Strange and exotic cityscapes. Desolate wilderness areas. Dogs that look like wookies. Flickr, one of the world's largest photo sharing services, sees it all. And, now, Flickr's image recognition technology can categorize more than 11 billion photos like these. And it does it automatically. It's called "Magic View." Magical deep learning! Buzzword attack!</div>

...

Cancel
...4 more comments...
LeopoldS on 21 Jul 15

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

<div class="cArrow"> </div><div class="cContentInner">and here comes my standard question: how can we use this for space? fast detection of natural disasters onboard?</div>

...

Cancel
Alexander Wittig on 22 Jul 15

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).

<div class="cArrow"> </div><div class="cContentInner">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).</div>

...

Cancel
Alexander Wittig on 23 Jul 15

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).

<div class="cArrow"> </div><div class="cContentInner">In fact, I think this could make a nice ACT project: counting seals from satellite imagery is an actual (and quite recent) thing: <a href="http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0092613" rel="nofollow" target="_blank">http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0092613</a> 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).</div>

...

Cancel
jcunha on 23 Jul 15

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.

<div class="cArrow"> </div><div class="cContentInner">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.</div>

...

Cancel
Paul N on 29 Jul 15

You're forgetting power usage. GPUs need 1000 hamster wheels worth of power while FPGAs can run on a potato. Since space applications are highly power limited, putting any kind of GPU monster in orbit or on a rover is failed idea from the start. Also in FPGAs if a gate burns out from radiation you can just reprogram around it. Looking for seals offline in high res images is indeed definitely a GPU task.... for now.

<div class="cArrow"> </div><div class="cContentInner">You're forgetting power usage. GPUs need 1000 hamster wheels worth of power while FPGAs can run on a potato. Since space applications are highly power limited, putting any kind of GPU monster in orbit or on a rover is failed idea from the start. Also in FPGAs if a gate burns out from radiation you can just reprogram around it. Looking for seals offline in high res images is indeed definitely a GPU task.... for now.</div>

...

Cancel
jcunha on 31 Jul 15

The discussion of how to make FPGA hardware acceleration solutions easier to use for the 'layman' is starting btw http://reconfigurablecomputing4themasses.net/.

<div class="cArrow"> </div><div class="cContentInner">The discussion of how to make FPGA hardware acceleration solutions easier to use for the 'layman' is starting btw <a href="http://reconfigurablecomputing4themasses.net/" rel="nofollow" target="_blank">http://reconfigurablecomputing4themasses.net/</a>.</div>

...

Cancel

Neurokernel - 4 views

neurokernel.github.io/index.html

BIO science INF

shared by Marcus Maertens on 01 Oct 14 - No Cached

Ma Ru liked it

Marcus Maertens on 01 Oct 14

A nice GPU-based framework that is basically an emulator of the brain of the fruit fly. If you need a fruit fly brain - here it comes!

<div class="cArrow"> </div><div class="cContentInner">A nice GPU-based framework that is basically an emulator of the brain of the fruit fly. If you need a fruit fly brain - here it comes!</div>

...

Cancel

NVIDIA Tesla - GPU Computing Solutions for HPC - 0 views

www.nvidia.com/...tesla_computing_solutions.html

supercomputing technology GPU

shared by ESA ACT on 24 Apr 09 - Cached

ESA ACT on 24 Apr 09

Graphic cards in the core of a supercomputer.

<div class="cArrow"> </div><div class="cContentInner">Graphic cards in the core of a supercomputer.</div>

...

Cancel

Twilight of the GPU - 1 views

arstechnica.com/...gpu-sweeney-interview.ars

technology forecast

shared by ESA ACT on 24 Apr 09 - Cached

ESA ACT on 24 Apr 09

Announcement of the end of an era.

<div class="cArrow"> </div><div class="cContentInner">Announcement of the end of an era.</div>

...

Cancel

NVIDIA Releases Open-Source GPU Kernel Modules - 0 views

https://developer.nvidia.com/blog/nvidia-releases-open-source-gpu-kernel-modules/

technology AI

started by eblazquez on 12 May 22 no follow-up yet

Convolutional networks start to rule the world! - 2 views

www.image-net.org/...results.html

computer vision AI neural networks

shared by Guido de Croon on 12 Jan 13 - No Cached

LeopoldS and Juxi Leitner liked it

Guido de Croon on 12 Jan 13

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 :)

<div class="cArrow"> </div><div class="cContentInner">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: <a href="http://www.idsia.ch/~juergen/vision.html" rel="nofollow" target="_blank">http://www.idsia.ch/~juergen/vision.html</a>) This should please Francisco :)</div>

...

Cancel
...1 more comment...
LeopoldS on 19 Jan 13

where is Francisco when one needs him ...

<div class="cArrow"> </div><div class="cContentInner">where is Francisco when one needs him ...</div>

...

Cancel
Dario Izzo on 22 Jan 13

...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

<div class="cArrow"> </div><div class="cContentInner">...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 <a href="http://www.sarjis.info/stripit/abstruse-goose/496/the_singularity_is_way_over_there.png" rel="nofollow" target="_blank">http://www.sarjis.info/stripit/abstruse-goose/496/the_singularity_is_way_over_there.png</a></div>

...

Cancel
johannessimon81 on 23 Jan 13

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..?

<div class="cArrow"> </div><div class="cContentInner">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..?</div>

...

Cancel

CPU and password strength - 4 views

www.zdnet.com/...ering-strong-passwords-useless

cryptography

shared by LeopoldS on 22 May 15 - No Cached

LeopoldS on 22 May 15

true?

<div class="cArrow"> </div><div class="cContentInner">true?</div>

...

Cancel
Paul N on 26 May 15

Isn't that why systems have a "wait for 15 minutes before trying again" after 3 or 5 wrong guesses? All the brute force in the word can't save you from real-life latency.

<div class="cArrow"> </div><div class="cContentInner">Isn't that why systems have a "wait for 15 minutes before trying again" after 3 or 5 wrong guesses? All the brute force in the word can't save you from real-life latency.</div>

...

Cancel
Ma Ru on 28 May 15

Oh, so you haven't heard about diceware yet? http://world.std.com/~reinhold/diceware.html And, of course, a related XKCD...

<div class="cArrow"> </div><div class="cContentInner">Oh, so you haven't heard about diceware yet? <a href="http://world.std.com/~reinhold/diceware.html" rel="nofollow" target="_blank">http://world.std.com/~reinhold/diceware.html</a> And, of course, a related XKCD...</div>

...

Cancel

CUDA-Enabled Apps: Measuring Mainstream GPU Performance : Help For the Rest of Us - Rev... - 0 views

www.tomshardware.com/...nvidia-cuda-gpgpu,2299.html

gpgpu cuda nvidia computer science

shared by Francesco Biscani on 19 May 09 - Cached

Francesco Biscani on 19 May 09

What will the name of the CUDA-enabled PaGMO be? CuDMO, CyGMO?

<div class="cArrow"> </div><div class="cContentInner">What will the name of the CUDA-enabled PaGMO be? CuDMO, CyGMO?</div>

...

Cancel

[1104.1824] Simulating Spiking Neural P systems without delays using GPUs - 4 views

arxiv.org/1104.1824

BIO

shared by Francesco Biscani on 12 Apr 11 - No Cached

Francesco Biscani on 12 Apr 11

Might be interesting for Giusy...

<div class="cArrow"> </div><div class="cContentInner">Might be interesting for Giusy...</div>

...

Cancel

Google Just Open Sourced the Artificial Intelligence Engine at the Heart of Its Online ... - 2 views

www.wired.com/...artificial-intelligence-engine

technology AI

shared by Daniel Hennes on 10 Nov 15 - No Cached

johannessimon81, Paul N, and jcunha liked it

Daniel Hennes on 10 Nov 15

TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. The flexible architecture allows you to deploy computation to one or more CPUs or GPUs in a desktop, server, or mobile device with a single API. TensorFlow was originally developed by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research, but the system is general enough to be applicable in a wide variety of other domains as well.

<div class="cArrow"> </div><div class="cContentInner">TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. The flexible architecture allows you to deploy computation to one or more CPUs or GPUs in a desktop, server, or mobile device with a single API. TensorFlow was originally developed by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research, but the system is general enough to be applicable in a wide variety of other domains as well.</div>

...

Cancel
Paul N on 11 Nov 15

And the interface even looks a bit less retarded than theano

<div class="cArrow"> </div><div class="cContentInner">And the interface even looks a bit less retarded than theano</div>

...

Cancel

GTC On-Demand Featured Talks | GPU Technology Conference - 3 views

on-demand-gtc.gputechconf.com/...on-demand-gtc.php

AI

shared by Marcus Maertens on 27 Nov 18 - No Cached

Marcus Maertens on 27 Nov 18

NVIDIA published around 154 talks focussed on AI from their conference this year...

<div class="cArrow"> </div><div class="cContentInner">NVIDIA published around 154 talks focussed on AI from their conference this year...</div>

...

Cancel

Facebook does Go - 3 views

research.fb.com/cebook-open-sources-elf-opengo

AI

shared by dharmeshtailor on 03 May 18 - No Cached

dharmeshtailor on 03 May 18

Took 2000 GPUs over 2 weeks to train :)

<div class="cArrow"> </div><div class="cContentInner">Took 2000 GPUs over 2 weeks to train :)</div>

...

Cancel

Introducing Triton: Open-Source GPU Programming for Neural Networks - 1 views

www.openai.com/triton

INF AI

shared by pablo_gomez on 28 Jul 21 - No Cached

pablo_gomez on 28 Jul 21

Might be of interest for torchquad and other projects

<div class="cArrow"> </div><div class="cContentInner">Might be of interest for torchquad and other projects</div>

...

Cancel

Group items tagged