A 2019 Guide for Automatic Speech Recognition
heartbeat.fritz.ai
A 2019 Guide for Automatic Speech Recognition
1–10 of 39 posts
Re: A 2019 Guide for Automatic Speech Recognition
#2Related guide from last week: A 2019 Guide to Speech Synthesis with Deep Learning (https://news.ycombinator.com/item?id=20819672)
Re: A 2019 Guide for Automatic Speech Recognition
#3Off topic: Why does that site not show the scrollbar? I use it to work out how long a post is...
Re: A 2019 Guide for Automatic Speech Recognition
#4So how can one actually use one of these systems? I'm familiar with pocketsphinx, where you download it and then run procketsphinx_continuous and it prints a transcription of the microphone. Or Julius, where you write a grammar and run it with that and it prints the output. The process for DeepSpeech seems to be along the lines of "get access to a machine with ten GPUs, find a huge dataset, train it, and then run that somehow" - are there compiled models that can actually be shipped as part of an application?
Re: A 2019 Guide for Automatic Speech Recognition
#5I'm a little confused about the title because the first paper is from 2014.
It's also too bad this doesn't mention any traditional HMM-based ASR techniques, as HMMs continue to be used on many SOTA systems, particularly those that can be reproduced publicly: https://github.com/syhw/wer_are_we
Re: A 2019 Guide for Automatic Speech Recognition
#6So how can one actually use one of these systems? I'm familiar with pocketsphinx, where you download it and then run procketsphinx_continuous and it prints a transcription of the microphone. Or Julius, where you write a grammar and run it with that and it prints the output. The process for DeepSpeech seems to be along the lines of "get access to a machine with ten GPUs, find a huge dataset, train it, and then run tha…
Verbio offers an ASR with built-in grammars, which is what you are asking for. Deepspeech is hot garbage in my experience.
Re: A 2019 Guide for Automatic Speech Recognition
#7Off topic: Why does that site not show the scrollbar? I use it to work out how long a post is...
It's a design decision that is gaining a lot of traction lately.
Re: A 2019 Guide for Automatic Speech Recognition
#8So how can one actually use one of these systems? I'm familiar with pocketsphinx, where you download it and then run procketsphinx_continuous and it prints a transcription of the microphone. Or Julius, where you write a grammar and run it with that and it prints the output. The process for DeepSpeech seems to be along the lines of "get access to a machine with ten GPUs, find a huge dataset, train it, and then run tha…
Obviously the big cloud players offer their own APIs and SDKs (for a price), but there are a few other solutions worth looking at.
Facebook has open sourced some pre-trained models: https://github.com/facebookresearch/wav2letter
Picovoice has some smaller, more efficient models capable of running on edge devices: https://github.com/Picovoice
Full ASR does require quite large models and datasets, but you don't need nearly that much power or data to fine-tune a model for your own domain.
Re: A 2019 Guide for Automatic Speech Recognition
#9I'm a little confused about the title because the first paper is from 2014. It's also too bad this doesn't mention any traditional HMM-based ASR techniques, as HMMs continue to be used on many SOTA systems, particularly those that can be reproduced publicly: https://github.com/syhw/wer_are_we
This.
The article quotes DeepSpeech, wavenets, LSTMs of all sorts; essentially, all the neural networks that scale terribly. DeepSpeech for example is pretty heavy and requires a decent GPU to get anywhere near a realtime factor of 1.
Meanwhile ASR through HMM's consistently hits realtime factors sub-1 and can run on small CPUs. e.g. the default models Kaldi ships with outperform DeepSpeech on a lot of modern examples, and are exponentially faster.
Moreover, training these HMM's is something that is feasible for a normal developer. Training newer models requires data of scale and quality (iirc Mozilla's models are trained on Common Speech which is an enormous crowd sourced dataset, and Google's wavenet models use an internal dataset of very high quality and quantity).
Until the models get more practically achievable, ASR for average people will probably continue to be dominated by Kaldi, Sphinx etc
Re: A 2019 Guide for Automatic Speech Recognition
#10I'm a little confused about the title because the first paper is from 2014. It's also too bad this doesn't mention any traditional HMM-based ASR techniques, as HMMs continue to be used on many SOTA systems, particularly those that can be reproduced publicly: https://github.com/syhw/wer_are_we
Yup. Pretty poor blogpost, basically just picking the most hyped up things.