Look around. We’re surrounded by technology that until recently seemed like it could only exist in the realm of science fiction. The most prominent example? Computers we can talk to. In 2020, it was estimated that there were as many as 4.2 billion voice assistant devices being used worldwide, and that number is expected to reach a whopping 8 billion by 2023. The launch of Google’s voice search in 2011, and Apple’s Siri shortly afterwards, signalled the beginning of a new era for this technology - an era that is now in full swing since Amazon’s Alexa joined the party. But the increased ubiquity and improved computing power of these devices doesn’t mean a free ticket to unlimited progress. There are still significant challenges: understanding natural speech, tackling error rates and earning trust, to name but a few. Learn how voice recognition works, and the challenges it faces, in this explainer video from iluli by Mike Lamb – explaining the ideas and innovations changing our lives. *New videos on the 1st of every month. Don't forget to like and subscribe:* https://bit.ly/subscribe_to_iluli *Or visit the iluli website:* https://www.iluli.eu _____ *Sources and Further Reading:* • _The machines that learned to listen:_ https://www.bbc.com/future/article/20170214-the-machines-that-learned-to-listen • _Digital Voice Assistants in Use to Triple to 8 Billion by 2023, Driven by Smart Home:_ https://www.juniperresearch.com/press/digital-voice-assistants-in-use-to-8-million-2023 • _How does speech recognition software work?_ https://www.explainthatstuff.com/voicerecognition.html _____ 0:00 Introduction 0:19 History of speech recognition 0:41 Recent progress 1:17 Spectogram 1:56 How it works 3:00 Challenges 3:47 Winograd Schema 4:07 Ethics and trust 4:24 Conclusion

• published: 01 May 2021
• views: 137661