You were recently invited to keynote for the AVIOS Israel Chapter’s annual conference. First, what is AVIOS and how did you become involved with the organization?
AVIOS (Applied Voice Input/Output Society) is a speech technology association founded in 1981 and with a membership consisting mainly of speech experts. I started in speech technology in 1984 and many of the others have been around for some time too. As it’s a small community, everyone pretty much knows everyone else.
With all the incredible technological innovations happening in the region, there must have been a great collection of speakers and attendee companies at the AVIOS Israel event. What are some of the key takeaways you brought with you from the show? Or some of the interesting side conversations that came up?
From the presentations that were in English (rather than Hebrew) and a few side discussions, three things stood out:
1. Speech technology tends to be concentrated in certain areas – Silicon Valley, Boston, Cambridge UK, etc., and it’s clear from the caliber of the audience that Israel is one of those places.
2. Speaker verification (voice biometrics that confirm a speaker’s identity) is drawing increasing interest and investment.
3. There is a resurgence of new R&D for voice message to text conversion. This is the most difficult problem in speech, but it is crossing the threshold from impossible to possible. (At the same time, success at SpinVox is proving that there is a market hungry for the technology.) We see a corresponding uptick in research in this area, including some new efforts in Israel.
Congratulations on presenting as the event’s keynote speaker this year! Tell us about your “Crossing Speech Technology Thresholds” presentation and why you chose that theme to share with this audience. And, what was the reaction to ‘Pedro the Voder’?
Pedro the Voder was the first electronic talking machine, demonstrated at the 1939 World’s Fair. I showed a bit on how it worked and played an old recording. The AVIOS audience was entertained, I think, but also realized how far we’ve come in 70 years. I mentioned the first network-based speech recognizer that understood only three words (1989) and analyzed progress in the 20 years since then.
The presentation showed that the market is eager to adopt speech technology as it becomes ready, but that just having an impressive prototype or even a successful trial does not qualify a service as “ready.”
I also related a few interesting events from the days when speech technology was still new which have never been reported or published. There have been some astounding breakthroughs over the past 25 years and each breakthrough has enabled a new set of applications. The latest breakthrough – converting recorded voice messages on any topic from any speaker into text – is triggering a new wave of services in voicemail-to-text, voice-to-SMS, apps for handheld devices, and call center automation.
Your talk included some predictions. What do you see happening in the next few years?
We can extrapolate from the past and look at the breakthroughs people are working on now to get a reasonable idea where things are headed. We see data becoming available everywhere and information becoming decentralized. Combined with our new ability to convert speech into text, we have now fully bridged the gap between data and voice, and we will see more of this available on handheld devices. Whereas standard voicemail has been on the decline, we see voicemail-to-text driving a dramatic upturn. We should also expect applications powered by voice-to-text to proliferate, including speech-to-SMS, blogging by voice, and voice powered web search, followed in a few years by language translation, intelligent concierge services, and automated systems that are as easy to talk to as live agents.
Sounds like an interesting presentation, how could I get a copy and learn more?
The paper is available here
And, the presentation is available here
What’s next for you and/or AVIOS?
On January 9th, I’m speaking at the Consumer Electronics Show, a massive conference in Las Vegas. My topic is “The New User Interface: Who Needs Buttons and Dials?” The next AVIOS event is the main annual conference, recently renamed to “Mobile Voice,” in San Francisco April 22-23.
Book Mark it-> del.icio.us | Reddit | Slashdot | Digg | Facebook | Technorati | Google | StumbleUpon | Window Live | Tailrank | Furl | Netscape | Yahoo | BlinkList