"Now you're going to try that and put it in action and see what it actually is like to use Windows Vista speech recognition."
"Speech recognition used to suck, but now it's pretty darn good."
"Speech to text allows users to compose text by speaking into their device."
"Speech recognition technology helps with pronunciation."
"Google has the best speech to text on the market."
"I can convert speech to text. I can translate from one language to another language."
"We've got the web speech API side of things right, so awesome stuff."
"We showed that on this very simple problem of vowel recognition."
"Speech API, this API has speech recognition and speech synthesis."
"Dragon NaturallySpeaking for those of you that are not computer buffs is a program that allows you to speak into your computer and it writes or types or displays what you say."
"The first wave of AI that focused on computer vision and speech recognition has achieved superhuman capabilities."
"AI technology is getting near flawless at hearing words spoken and then inputting it to text."
"Sequence models have transformed speech recognition, natural language processing, and other areas."
"One of the most exciting developments with sequence-to-sequence models has been the rise of very accurate speech recognition."
"The best commercial systems are now trained on over ten thousand hours, and sometimes over a hundred thousand hours of audio."
"This is sometimes called the fundamental equation of speech recognition."
"For the uninitiated, what is an ASR system? It's one which accurately translates spoken utterances into text."
"So the language model is very crucial because it can be used to disambiguate between similar acoustics."
"Speech recognition... it's definitely not a solved problem and we certainly have plenty of work we can keep doing to try to improve on this."
"I fell in love with speech recognition."
"Speech to text: unreal on this phone, absolutely incredible."
"This phone right here has the best text to speech on the planet."
"We'll use a special loss function called CTC."
"Speech recognition integrates the probability of what is being said with what was said previously."
"The robot speaks Arabic and English, successfully recognizes and interprets speech."
"So, flashback to some years ago, the motivating problem for speech recognition was this thing called the cocktail party problem."
"The ultimate goal of solving speech recognition is to improve people's lives and enable exciting products."
"Deep learning is really making the first steps to building a state-of-the-art speech engine easier than they've ever been."
"Speech recognition is at a place right now where it's becoming good enough to enable really exciting applications."
"These kind of approaches have reduced the word error rate in the speech recognition system by more than 30 percent."
"Speech recognition can be tackled with HMMs."
"We thought it was the first working speaker independent continuous speech recognition and we were very proud of it."
"Speech recognition is a fascinating field where we can teach computers the ability to recognize spoken words."
"Speech recognizer converts audio to text."
"The potential of speech recognition and technology... it has potential not just to help... it has potential to also uplift communities."
"Real-time speech recognition is not magic anymore."
"Machine learning has also significantly improved speech recognition."
"People are using speech to code in real life today... it's not like this is a thing that may happen in the future, it's happening right now."
"Deep learning is much more than just image processing; one such application is speech and natural language processing."
"We have very good speech recognizers, but they are all based on machine learning."
"Recurrent neural networks are pretty good for doing NLP or speech recognition and so on."
"Speech recognition and image recognition have never been this good."
"The key to successful speech recognition will be that when you correct Siri, Siri will get better."
"It's a very simple tool that uses the Google speech recognition technology or Google AI to create very accurate transcriptions."
"You don't need to use any other speech-to-text API; it is inbuilt to MRTK."
"The speech recognition technology helps you improve your pronunciation and accent."
"Deep speech is the latest and greatest research coming out of Baidu to try and tackle these sorts of problems."
"If you want a computer to understand words very clearly, you need to really say clearly."
"Speech recognition at Google is getting so good that Nino has replaced a lot of his typing with speaking."
"We actually want to make personalized model that understand how you speak, your pace of speech, the words you say most commonly, accents, et cetera."
"We'll create a bit that recognizes speech from your microphone, understands those commands as text."