Chapter
Overcoming Far Field Speech Recognition with Amazon Echo
48:23 - 1:01:18 (12:55)
Amazon Echo's far field speech recognition was made possible with a combination of large data sets, deep learning progress, and infinite GPUs on AWS, which solved the problem of detecting the right mentions of Alexa's address to the device versus other general statements.
Clips
The director of machine learning at Amazon believes there's no negative impact for customers if Alexa could listen all the time to their conversations, with their permission, to learn more about them and enhance their experience with the device.
48:23 - 52:24 (04:01)
Summary
The director of machine learning at Amazon believes there's no negative impact for customers if Alexa could listen all the time to their conversations, with their permission, to learn more about them and enhance their experience with the device. He says he worries how sensitive people are about their data relative to how empowering it could be for the devices around them.
ChapterOvercoming Far Field Speech Recognition with Amazon Echo
EpisodeRohit Prasad: Amazon Alexa and Conversational AI
PodcastLex Fridman Podcast
The problem of accurately detecting the wake word "Alexa" from audio that may come from far away or amidst noise and other conversations is a challenging one for large vocabulary speech recognition systems.
52:24 - 59:03 (06:38)
Summary
The problem of accurately detecting the wake word "Alexa" from audio that may come from far away or amidst noise and other conversations is a challenging one for large vocabulary speech recognition systems.
ChapterOvercoming Far Field Speech Recognition with Amazon Echo
EpisodeRohit Prasad: Amazon Alexa and Conversational AI
PodcastLex Fridman Podcast
Amazon's ability to solve far field speech recognition for its Echo device was made possible by the combination of deep learning progress, the availability of near infinite GPUs on AWS and large scale data.
59:03 - 1:01:18 (02:15)
Summary
Amazon's ability to solve far field speech recognition for its Echo device was made possible by the combination of deep learning progress, the availability of near infinite GPUs on AWS and large scale data. While not perfect, Amazon's speech recognition capabilities are well-suited for household settings.