Awni Hannun, Carl Case, Jared Casper
Bryan Catanzaro, Greg Diamos, Erich Elsen, Ryan Prenger, Sanjeev Satheesh, Shubho Sengupta, Adam Coates, Andrew Y. Ng
2014 · arXiv
We present a state-of-the-art speech recognition system developed using end-to-end deep learning. Our architecture is significantly simpler than traditional speech systems, which rely on laboriously engineered processing pipelines; these traditional systems also tend to perform poorly when used in noisy environments. In contrast, our system does not need hand-designed components to model background noise, reverberation, or speaker variation, but instead directly learns a function that is robust to such effects. We do not need a phoneme dictionary, nor even the concept of a "phoneme." Key to our approach is a well-optimized RNN training system that uses multiple GPUs, as well as a set of novel data synthesis techniques that allow us to efficiently obtain a large amount of varied data for training. Our system, called Deep Speech, outperforms previously published results on the widely studied Switchboard Hub5'00, achieving 16.0% error on the full test set. Deep Speech also handles challenging noisy environments better than widely used, state-of-the-art commercial speech systems.
Metadata: arXiv. Abstracts retain their original authors’ rights.
Original HN comments, with labels from simple text rules. A label describes the whole comment and may miss context or irony. Read the discussion to make up your own mind. About the labels
5 selected excerpts · 5 total mentions
Mozilla announced Deep Speech[1] around the same time as Common Voice. Mozilla Deep Speech is an open source speech recognition engine, based upon Baidu's Deep Speech research paper[2]. Unsurprisingly, Deep Speech requires a corpus such as... Common Voice. [1] https://github.com/mozilla/DeepSpeech [2] https://arxiv.org/abs/1412.5567
Added an enhancement request, issue 1064[1], to github, asking for the clients to support longer audio clips. I can't promise when we'll get to it, as from now until new year is a bit of a wash. I don't know of any detailed comparisons of commercial solutions. However, with respect to pure word error rate, the article[2] does a comparison of several engines as of circa 2015. [1] https://github.com/mozilla/DeepSpeech/issues/1064 [2] https://arxiv.org/abs/1412.5567
Deepgram just posted an open sourced Kur Speech Recognition example based on Deepspeech (https://arxiv.org/abs/1412.5567) on our blog. We'd love to help people get it running and spread the DNN speech love.
> Of course, OCR and speech recognition are somewhat different fields, but if anyone is an expert for Chinese language, then it should be Baidu. It's not about being experts in Mandarin. The basis of their approach is that it doesn't encompass any expert design. It's an end-to-end deep learning approach. From the article: > Our system is different than that system in that it’s more what we call end-to-end. Rather than having a lot of human-engineered components that have been developed over decades of speech research — by looking at the system and saying what features are important or which phonemes the model should predict — we just have some input data, which is an…
audio .WAV file on which we do very little pre-processing. And then we have a big, deep neural network that outputs directly to characters. We give it enough data that it’s able to learn what’s relevant from the input to correctly transcribe the output, with as little human intervention as possible. > One thing that’s pleasantly surprising to us is that we had to do very little changing to it — other than scaling it and giving it the right data — to make this system we showed in December that worked really well on English work remarkably well in Chinese, as well. So they've quickly trained their Deep Speech engine [1] to process Mandarin after first training it to transcribe English, without injecting specific language expertise into the engine. Finally, I strongly doubt the OCR and speech recognition teams are the same. I don't know about the OCR team but their speech recognition team is based in California [2] and includes Andrew Ng and Awni Hannun from Stanford University. [1] http://arxiv.org/abs/1412.5567 [2] http://usa.baidu.com/deep-speech-lessons-from-deep-learning/
Actually, state of the art speech recognition has switched over to having a Recursive Neural Network directly run over the audio input. Take a look at the paper at http://arxiv.org/abs/1412.5567 and http://usa.baidu.com/deep-speech-lessons-from-deep-learning/ However combining learning features with other systems is a very powerful approach and combining SVM's on top of the learned features of a Neural Network I would say is common. I personally am more interested in approaches like Deep Fried Convnets (http://arxiv.org/abs/1412.7149) that combine kernel methods as part of the Neural Networks themselves.
More from the same topics.
Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan +11
Artificial intelligence (AI) researchers have been developing and refining large language models (LLMs) that exhibit remarkable capabilities across a variety of domains and tasks, challenging our understanding of learning and cognition.
Tom B. Brown, Benjamin Mann, Nick Ryder +28
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task.
Chiyuan Zhang, Samy Bengio, Moritz Hardt +2
Despite their massive size, successful deep artificial neural networks can exhibit a remarkably small difference between training and test performance.
A. Pluchino. A. E. Biondo, A. Rapisarda
The largely dominant meritocratic paradigm of highly competitive Western cultures is rooted on the belief that success is due mainly, if not exclusively, to personal qualities such as talent, intelligence, skills, efforts or risk taking.