> Neural networks are good at adaptation, but useless at forming concepts about how the data is structured. For example: in video we do motion compensation, because we know video captures motion since objects move in physical reality. A neural network would have to do the same in order to get the same compression levels.
You're not supposed to. Image generation and modeling scene dynamics is a hard task, and thumbnail scale is what we're at at the moment. Nevertheless, those and other papers do demonstrate that NNs are perfectly capable of learning 'objectness' from video footage (to which we can add GAN papers showing that GANS learn, based on static 2D images, 3D movement). More directly connected to image codecs, there's work on NN prediction of optical flow and inpainting.
Many NNs can be compressed considerably without losing much performance. The runtime of RNNs is more concerning, as is whether anyone wants to move to a new image format, but it's still interesting pure research in terms of learning to predict image contents. It's a kind of unsupervised learning, which is one of the big outstanding questions for NNs at the moment.
Not sure there's anything to explain. Waking up and then going back to sleep is going to help with inducing lucid dreams, supplement-assisted or not, that's the point of WILD.
You're right - a $3k k80 is generally inferior for most DNN applications than a $1k Titan X. The primary reasons that big companies use K80s have to do with achieving high computational density, and licensing issues, more than the performance of a single GPU. Sticking 8 Titan X boards in a machine is a bummer job if you want to pack them closely together. But for academic researchers, a quad Titan X box is pretty solid and quite affordable.
A quad Titan X is still $4k for the GPUs alone, and was only possible in the past few months - people might have wanted to get stuff done in the years in between the last generation and the current generation...
You compare it to the general population. In disparate impact lawsuits, this is the 'applicant' vs 'pool' distinction; the idea is to make it possible to sue even when the discrimination 'barriers' operate before the formal application process, but of course, you wind up in Orwellian games where the necessary number of minority potential-employees simply do not exist and the employer is forced to provide a devil's proof to show they weren't being discriminatory and they argue over what 'proxies' are appropriate... Ward Cove and https://en.wikipedia.org/wiki/Hazelwood_School_District_v._U...
The Democratic party targeting university students, and the Republican party trying voter-suppression efforts aimed at students (forbidding voting at the university where one lives, requiring photo IDs & banning college IDs as photo IDs) are nothing new.
That raises as many questions as it answers. Why FPGAs and not GPUs which can run just about any deep neural network but usually faster and more efficiently?
Sure, GPUs deliver impressive raw performance. To be useful, the task must benefit from massively parallel hardware. GPU hardware works fantastic for shading polygons, training neural networks, or raytracing. For compression and encryption algorithms however, GPUs aren’t terribly good.
Another reason is while a GPU delivers impressive bandwidth on parallel-friendly workloads, it’s usually possible to achieve lower latencies with FPGA. An FPGA doesn’t decode any instructions, and its computing modules exchange data directly.
GPUs give you great throughput, but they're expensive, eat a lot of power, only work well in batch and aren't tuned to prediction (eg poor INT8 performance and too much VRAM).
GPUs have worse performance per watt than a tuned FPGA. Some newer FPGAs can have 400 megabits of on chip RAM - that's huge, significantly larger than the 128-256k cache typically available on chip for a GPU that turns into big energy savings.
While I dont have any external publications addressing this general claim, this is taken from my current and past experiences with internal studies focused on neural networks implemented on the TX1, other GPU, custom ASICs, and FPGA approaches. In terms of power efficiency it generally goes ASIC > FPGA > GPU > CPU. If you're doing just fp32 BLAS it's hard to beat a GPU, but it turns out many problems have features that you can optimize for.
The TX1 power consumption including DRAM and other subsystems peaks 20-30W. Typical usage is 10-15W if you're running anything useful.
That 1 TFLOP counts a FMA instruction as 2 flops - while accurate and useful for say dot products - for other workloads the throughput will be half of this number.
As an example of an FPGA performing significantly better than the TX1 is DeepPhi [0].
While not the TX1 vs FPGA result you want, this is very close. For example they aren't using the latest FPGA or GPU, and are not using TensorRT on the GPU and on the FPGA side they are using fatty 16-bit weights on an older FPGA rather than newer stuff you can do with lower precision (which improves the efficiency of the FPGA having more high speed RAM collocated with computation vs GPU which is primarily off-chip).
Thanks, but TK1 is using FP32 weights, as opposed to FP16 on FPGA. If you double the GOP/s number for TK1 to account for that, you will end up with pretty much identical performance, and the paper claims they both consume ~9W.
I'm not saying you're wrong, just that to make a convincing claim that FPGAs are more power efficient than GPUs, one needs to do an apples to apples comparison.
And of course, let's not forget about price: Zynq ZC706 board is what, over $6k? And Jetson TK1 was what when released, $300? If you need to deploy a thousand of these chips in your datacenter, to save a million per year on power, you will need several years to break even, and by that time, you will probably need to upgrade.
It just seems that GPUs are a better deal currently, with or without looking at power efficiency.
Neural networks know that objects move. See for example https://arxiv.org/abs/1511.05440 http://web.mit.edu/vondrick/tinyvideo/