DeepSpeech: Scaling up end-to-end speech recognition
1–10 of 23 posts
Re: DeepSpeech: Scaling up end-to-end speech recognition
#2Thought it best to post the arXiv link, but there's some press coverage as well:
- https://gigaom.com/2014/12/18/baidu-claims-deep-learning-bre... - http://www.forbes.com/sites/roberthof/2014/12/18/baidu-annou...
Re: DeepSpeech: Scaling up end-to-end speech recognition
#3Thought it best to post the arXiv link, but there's some press coverage as well: - https://gigaom.com/2014/12/18/baidu-claims-deep-learning-bre... - http://www.forbes.com/sites/roberthof/2014/12/18/baidu-annou...
I should add that I had the opportunity to work on this project and am happy to answer questions.
Re: DeepSpeech: Scaling up end-to-end speech recognition
#4This is very fast progress from Baidu's Silicon Valley AI lab! Andrew Ng only joined Baidu in May, and (nearly?) all of the co-authors of this paper have joined him since then:
http://www.technologyreview.com/news/527301/chinese-search-g...
Congrats to Carl, Sanjeev, Andrew, and the others.
Re: DeepSpeech: Scaling up end-to-end speech recognition
#5Thought it best to post the arXiv link, but there's some press coverage as well: - https://gigaom.com/2014/12/18/baidu-claims-deep-learning-bre... - http://www.forbes.com/sites/roberthof/2014/12/18/baidu-annou...
I should add that I had the opportunity to work on this project and am happy to answer questions.
Why 5 hidden layers? Why are the first 3 non-recurrent? How did you decide how wide to make the internal layers? Are there some organizing principles behind these design decisions, or is it just trial and error?
Re: DeepSpeech: Scaling up end-to-end speech recognition
#6Earlier quoted context omitted.
I should add that I had the opportunity to work on this project and am happy to answer questions.
Why 5 hidden layers? Why are the first 3 non-recurrent? How did you decide how wide to make the internal layers? Are there some organizing principles behind these design decisions, or is it just trial and error?
As in many things, it's a combination of both. For example:
- We wanted no more than one recurrent layer, as it's a big bottleneck to parallelization.
- The recurrent layer should go "higher" in the network, as it's more effective at propagating long-range context when using the network's learned feature representation than using raw input values.
Other decisions are guided by a combination of trial+error and intuition. We started on much smaller datasets which can give you a feel for the bias/variance tradeoff as a function of the number of layers, the layer sizes, and other hyperparameters.
Re: DeepSpeech: Scaling up end-to-end speech recognition
#7To put it in perspective, my team in IBM Watson has already published better numbers (10.4% WER vs 13.1% WER for Baidu) on the SWB dataset. We haven't run our model on the CH part so we can't compare on the full test set. Paper here: http://www.mirlab.org/conference_papers/International_Confer....
Re: DeepSpeech: Scaling up end-to-end speech recognition
#8Thought it best to post the arXiv link, but there's some press coverage as well: - https://gigaom.com/2014/12/18/baidu-claims-deep-learning-bre... - http://www.forbes.com/sites/roberthof/2014/12/18/baidu-annou...
I should add that I had the opportunity to work on this project and am happy to answer questions.
How much latency does the system have in the best/worst and average case? And is your implementation public?
Re: DeepSpeech: Scaling up end-to-end speech recognition
#9To put it in perspective, my team in IBM Watson has already published better numbers (10.4% WER vs 13.1% WER for Baidu) on the SWB dataset. We haven't run our model on the CH part so we can't compare on the full test set. Paper here: http://www.mirlab.org/conference_papers/International_Confer... .
Hi Jerome, those are great results! We got an email this morning from someone else on the Watson team pointing out that we didn't include the latest IBM number -- we'll be sure to update the results in the next version of the paper (three cheers for arXiv).
Of course, we openly say in the paper that we don't have the best result on easy subset of Hub5'00 (we had it as 11.5%). We're more interested in advancing the state of the art on challenging, noisy, varied speech. Of course we'll be working to push the SWB number down too :)
Re: DeepSpeech: Scaling up end-to-end speech recognition
#10Thought it best to post the arXiv link, but there's some press coverage as well: - https://gigaom.com/2014/12/18/baidu-claims-deep-learning-bre... - http://www.forbes.com/sites/roberthof/2014/12/18/baidu-annou...
I should add that I had the opportunity to work on this project and am happy to answer questions.
Any chance of releasing the training data you used? Also what are the plans with DeepSpeech? Just for use by baidu or will it be released as open source or a developer api service?