News for hackers. Stuff that matters.
Re:Speech Recognition Is Not Solved by gok on Tuesday October 24, 2017 @02:37PM (#15544052)
The TPU architecture isn't that weird...it's basically a hardware implementation of matrix multiplication. It also isn't a silver bullet for ASR, where neural networks are usually only used for a part of the recognition process.
Re:Speech Recognition Is Not Solved by dragontamer on Tuesday October 24, 2017 @04:12PM (#15544857)
> The TPU architecture isn't that weird

Its "weird" in the ways that matter: there's no commodity hardware in existence that replicates what a TPU does. The only place to get TPUs is through Google's cloud services.

CPUs are basically Von Neumann Architecture. GPUs (NVidia and AMD) are basically SIMD / SIMT systems.

Google's TPU is just something dramatically different, optimized yes for Matrix Multiplication, but its not something you can buy and use offline.

Re:Speech Recognition Is Not Solved by gok on Tuesday October 24, 2017 @05:06PM (#15545255)
It's not that dissimilar in architecture or performance to the Tensor Cores in Volta, which you can buy soon.
Re:Speech Recognition Is Not Solved by dragontamer on Tuesday October 24, 2017 @05:19PM (#15545326)
I wasn't aware of Tensor Cores in Volta.

I'll look into them, thanks!

Re:Speech Recognition Is Not Solved by Cacti on Tuesday October 24, 2017 @04:26PM (#15544951)
> but its not something you can buy and use offline.

But you will. The entire point is to put this in a phone, so you can distribute a trained neural net in a way that people can actually use without a desktop and $500-$4,000 GPU.

Re:Speech Recognition Is Not Solved by dragontamer on Tuesday October 24, 2017 @04:59PM (#15545212)
> But you will. The entire point is to put this in a phone, so you can distribute a trained neural net in a way that people can actually use without a desktop and $500-$4,000 GPU.

As far as I can tell, they put a microphone on your phone and then relay your voice to Google's servers for analysis.

Or Amazon's servers, in the case of the Echo.

I don't see any near-term future where Google's TPUs become widely available for consumers: be it on a phone or desktop. And I'm not aware of any product from the major hardware manufacturers that even attempt to replicate Google's TPU architecture.

NVidia and AMD are sorta going the opposite direction: they're making their GPUs more and more flexible (which will be useful in a wider variety of problems), while Google's TPUs specialize further and further into low-precision matrix multiplications.

Re:Speech Recognition Is Not Solved by shpx on Tuesday October 24, 2017 @05:42PM (#15545490)
Neural nets take large GPUs (or TPUs) to train. Realtime inference on CPUs has been possible since forever.

Also, I just turned on airplane mode and google assistant recognized my voice.

Re:Speech Recognition Is Not Solved by StillBored on Tuesday October 24, 2017 @04:53PM (#15545160)
Is that the point? I ask, because the "weird" in the TPU is mostly its scale. Its not like you can't do matrix multiplies with the vector units on a CPU or with a GPU. Its really the scale, by that I mean its more elements than what you get with existing hardware, but its also lower precision, and appears less flexible, and is bolted to a heavyweight memory subsystem.

So, in that regard its no more "weird" than other common accelerator/coprocessors for things like compression.

So, in the end, what would show up in a phone doesn't really look anything like a TPU. I would maybe expect a lightweight piece of matrix acceleration hardware, which due to power constraints isn't going to be able to match what a "desktop" level FPGA or GPU is capable of much less a full blown TPU.

Re:Speech Recognition Is Not Solved by jrkdgdjenr on Tuesday October 24, 2017 @07:39PM (#15546150)
If this were true i’d expect to see some effort to open and standardize the hardware. Otherwise what’s the point?
Re:Speech Recognition Is Not Solved by IshKebab on Tuesday October 24, 2017 @03:28PM (#15544492)
Neural networks are used for nearly all of ASR now. Last I heard only the spectral components were still calculated not using a neural net and the text-to-speech is now entirely neural network (i.e. you feed text in and get audio samples out). I'd be surprised if they don't do that for ASR too soon if they haven't already.
Re:Speech Recognition Is Not Solved by adrianbg on Tuesday October 24, 2017 @03:54PM (#15544706)
Although some models are end-to-end neural nets, most of the ones in production (and all of the ones that get state of the art results) only use a neural net for one part of the process. Lots of people are as surprised as you, but that's the way it is.

Edit: I should say that in state of the art results there tend to be multiple components, including multiple neural nets and the tricky "decode graph" that gok and I are talking about. These are trained separately then get stuck together, as opposed to being trained in an end-to-end fashion.

Re:Speech Recognition Is Not Solved by woodson on Tuesday October 24, 2017 @05:28PM (#15545400)
Separating acoustic model and decoding graph search makes sense since you would need a huge amount of (correctly!) transcribed speech for training. See, for example, this paper by Google [1], where they used 125,000 hours (after filtering out the badly transcribed ones from the original 500,000 hours of transcribed speech) for training an end-to-end acoustic-to-word model. Good "old-school" DNN acoustic models can already be trained with orders of magnitude less training data (hundreds to thousands of hours).

[1] https://arxiv.org/abs/1610.09975

Re:Speech Recognition Is Not Solved by adrianbg on Tuesday October 24, 2017 @08:05PM (#15546274)
Yes, exactly. I do wonder whether a similarly good end-to-end system could be trained by constraining the alignments as I've seen done in some papers.
Re:Speech Recognition Is Not Solved by computerex on Wednesday October 25, 2017 @09:27AM (#15549883)
AFAIK state of the art models are hybrid of HMM/GMM and CNN for phoneme classification. There are exotic CTC/RNN based architectures for end-to-end recognition but they aren't state of the art.