Voice Technologies are becoming ubiquitous today but there was a time when the only option was to let our fingers do the talking on keyboards. User interfaces have come a long way so that recent technological breakthroughs have allowed us to effectively “talk” to our technology to get things done. In the era of voice assistants people are using voice naturally to get directions, find recipes, listen to music, hear stories, play games, relax and more. Thanks to machine learning, Voice recognition is now accurate enough to make it a powerful option to materialize clever voice assistants besides a big range of other tasks, but also Voice generation has made a technological leap forward by incorporating neural networks to achieve unprecedented quality in artificial speech. Are we at an inflection point in voice technologies? What’s is powering them at scale? Why is this all happening now? In this talk we’ll go on a trip together starting with the first glimpses of voice tech in the industry, through the meaningful scientific research and technological evolution that, in spite of running into multiple dead ends, gave us things like Alexa or Siri. The topics I’ll cover: * Learn about the technological milestones that brought us here, the voice era * A look at the fundamental technologies behind language understanding (ASR, NLU/NLP, NTTS, Cloud computing) * A peek at the infrastructure requirements at scale that took the technology to mass adoption * A glimpse at how small problems in the space are being solved with the help of AI to orchestrate everything into intelligent voice assistants * How voice technologies are now going outside the living run into on-the-go scenarios like the phone and the car and embedded in home appliances. Finally we’ll take a peek at some hints of what’s to come in the near future as Voice tech relies even more on machine learning techniques to become more natural, more intelligent, more knowledgeable and more context aware. Being part of the Alexa team I have a privileged seat to see and influence how Voice Technologies are evolving fueled by the highest adoption rate of a new technology that world has ever seen. And this time I want to share some knowledge and part of the excitement that these technologies are causing in many of us since we’re the #VoiceFirst generation! Here are some of the topics I will cover in more detail: – How a voice assistant works behind the curtain – When does a voice assistant wakes up – How do voice assistants work in noisy environments – How your intention is inferred when you talk – How is context kept in a conversation – Adapting dynamically to the speaker language – How self correction is used to dispense with human intervention – How can voice assistants go beyond voice detection – Emotions in a voice interaction – How can specific voices be modeled such as a celebrity voice – How latent goal detection triggers a whole new experience in voice assistants.
