Hi all,

Has anyone tried Siri on the iPhone 4S? If so, how does it work? That is, does it have an exceptionally large command vocabulary? I am very intreagued by the natural way that people are able to talk to it and get answers.

 

Pranav

-------------------------

Pranav,

Siri is Nuance technology. However, it does not use a large vocabulary and uses constrained grammars. What you see on TV is what you get in a perfect world. Canned demos always work. In the real world, it is very effective, but it's far from perfect. You can get some bizarre results depending upon what you're asking for and how you phrase it. What it does have is a very good link in terms of interpreting what you're asking and the available responses. Makes it look like it's very sophisticated. However, it is still a first-generation application. In short, what you see is not always what you get, good as it is.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, we make it better!

Be careful what you wish for.  You may find that what you get is not what you expect, or what you want. - Aesop (620 BC - 700 BC modern interpretation)

-------------------------

My understanding is that Siri uses Nuance technology for speech recognition only and that Siri could use other speech recognition engines, if it so chose.

-------------------------

Matt,

True, but who are they going to use? No one else has as good a mobile speech recognition technology compared to Nuance. There are a few other Apple apps for the iPhone and all of them use Nuance speech technology. For example, I'm sure that you have seen the ads for language translation software on the iPhone. This also uses Nuance speech technology (mobile). From what I know, Apple tried other engines and found them significantly wanting in reliability and performance. Besides, Apple and Nuance are joined at the hip Re: mobile speech technology. Take a look at some of the older Apple and Nuance press releases.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, we make it better!

"Aiming for the moon and missing it is better than aiming for the ditch and hitting it."   - Author Unknown

-------------------------

Chuck, I don't know who else's technology could be used, but I am old enough to remember when Altavista was the dominant search engine.  I wouldn't assume that Nuance will always be the best source.  Here's what the co-founder of Siri said:

http://9to5mac.com/2011/10/03/co-founder-of-siri-assistant-is-a-world-changing-event-interview/

-------------------------

Matt,

That interview is interesting but there is some misleading information there. First of all, all of the speech technology that ScanSoft inherited from L&H, which includes mobile technology at L&H was working on it time, was acquired well before ScanSoft and Nuance merged. So, Nuance never had the technology regardless of who spawned them originally. All of the technology that is currently owned and patented by Nuance came from the acquisition of the L&H resources following the Chapter 11 debacle in December of 2001. Nuance was acquired until 2002/2003. So the information in the interview is somewhat misleading.

On the other hand, I don't dispute that Siri could possibly incorporate any speech engine, or the statement that the speech recognition or speech engine isn't that important. However, which still in use because that is the primary input used by the application. Once Siri understands what is being asked for, it would be quite obvious that the search engine is the critical component. Nevertheless, you have to have a speech interface to get there if you're going to talk to it.

Regardless, any replacement for the current Nuance speech engine will have to navigate around the myriad of patents owned by Nuance, and numerous patents pending. It ain't going to be that easy.

So, be careful what you read, particularly in this case, because it can be misleading.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, we make it better!

"The least questioned assumptions are often the most questionable." -- Paul Broca

-------------------------

Chuck, so what was the chapter 11 disaster exactly???

R

Quote:

what was the chapter 11 disaster exactly???

This is how we remember it: L & H artificially pumped up the value of their stock by approximately 33% shortly before purchasing NaturallySpeaking from DragonSystems. Unfortunately, Jim and Janet Baker, the former owners of NaturallySpeaking, apparently had no problem with L & H's claim of selling $300 million worth of L & H Voice Express (their speech recognition product) to North Korea; a country that doesn't have electricity, let alone computers. If our remark seems a bit cutting, we specifically warned Janet Baker that this deal didn't smell right but we were quickly informed that we were smalltime resellers that didn't understand the art of major finance. Although Janet's remark was no doubt correct, we found it difficult to understand how they could overlook the alleged $300 million North Korean sale. We believe that the Bakers were so relieved that someone would want to purchase their company after losing $40 million, over the course of 2 years; they were apparently willing to grasp at L & H's offer to save their sinking ship and only pretended to examine the L & H books. In the end, it was the Bakers who paid the ultimate price for ignoring everyone's financial advice.

The Bakers were accustomed to selling speech recognition to the military when toilet seats were selling for $220 apiece and hammers were selling for $80. They were completely unaccustomed to selling to the public but managed to hang on when they released NaturallySpeaking because it made Product of the Year 2 years running at the Nam show in Las Vegas. This gave DragonSystems free advertising but when they didn't make the Product of the Year on the 3rd year, they hemorrhaged until they were bought out.

Although the former president of L & H (we no longer remember his name) and the 2 owners (Lernout & Hauspie) were tried in Belgium, they didn't get much more than a slap on the wrist. It's rumored that the president walked away with $50 million and that the owners each walked away with $100 million. If we sound a little bitter, we were on the cusp of making a large sale to Vanderbilt University in Nashville when Vanderbilt decided that NaturallySpeaking was simply too big a risk and decided to go with the IBM ViaVoice product and we know how that turned out.

-------------------------

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012
Click SpeechStart+ to Download a 15 Day Evaluation of SpeechStart+

In terms of alternatives to Nuance, what else is out there?

Just out of curiosity, what exactly has happened to IBM's speech recognition technology?  Are they still developing it or have they repackaged it in some other commercial form?

What about Vlingo?  Whose technology is behind that?

Where is WSR, Windows speech recognition going?

The (very) senior IBM exec who sponsored Watson was also the earlier the sponsor of IBM's voice technology R&D which included call center technology as well as ViaVoice. All that IBM IP is licensed (and some sold) to Nuance.

The IBM portfolio is an added huge wall of lawyers (plus IP) added to Nuance. While I know the folks at IBM, I don't know the details of the IP agreement. However, it would be very typical to have a joint "must defend" clause which would give any potential competitor's corporate council severe heartburn after Nuance presented to him the "license from us or else" proposal.

Again there are tons of press releases about IBM and Nuance.

Phil Schaadt

-------------------------
DNS12 Pro BM V Large Vocabulary plus KB or Voice Computer running Win7 64 bit machines with i7-2640M to i7-3960x and i7-3770K processors plus a Sony VAIO Windows 8 machine. DBX Audio Gate with Allen&Heath mixer/USB Audio; Andrea PureAudio USB  usually with Airline 77 or Audio-Technica but also Sennheiser MD431 II, theBoom, et. al.

One of the most earth shattering events in the past decade, in my own opinion, was the victory of Watson in the game of Jeopardy over the 2 top human contestants.

I think that Siri is in some ways an attempt to apply similar technology, taking natural language queries and allowing them to be used in a natural way.  In other words, mimicking the actions of a human assistant.

I believe that Siri is still considered to be "beta software."

Hi Chuck and all,

 

Thanks for the answers. This thread is becoming more interesting every post. <chuckle

 Pranav

-------------------------

Lunis has the basics of what happened pretty well explained. However, it goes much deeper than that because Yo and Pol (Lernout & Hauspie) were actually pulling similar stunts in Belgium by supporting the creation of pseudo-companies for similar purposes. However, it was the Korean fiasco that was discovered by a reporter from the Wall Street Journal that started the fall from grace. Just prior to that, I was working for a company called TikSoft. The CEO of TikSoft and the CEO of L&H were old friends and they decided to merge both companies because TikSoft was distributing a product called Now You're Talking Deluxe that was competing with L&H VoiceXpress using the L&H speech engine. They both thought that merging would be better for both. After the merger, I became an employee of L&H and as a result of my friendship with the CEO of TikSoft, was privy to a lot of what went on that eventually brought down the house of cards. The details are more or less irrelevant at this point. However, suffice it to say that L&H was the first Enron. The end result was that the foundation of L&H was rocked so hard that even the efforts of some excellent interim CEOs failed.

The only detail that I would add to this is that the CEO of L&H at the time (Gaston Bastiens) and Yo & Pol were eventually convicted of criminal offenses in Belgium as a result of their indiscretions and ended up serving four years in jail over there. They may or may not have gotten away with a significant amount of money (i.e., no one is quite privy to the total dollars and cents on this issue), but part of the sentencing was they were required to pay restitution for their crimes and didn't walk away with as much as one would think. Needless to say, the whole thing is water over the dam now. However, they would've gotten away with much more if it weren't for the fact that Belgian law is significantly tougher than US law as regards what they did.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make speech recognition, we make it better!

Knowing a great deal is not the same as being smart; intelligence is not information alone but also judgment, the manner in which information is collected and used. -  Dr. Carl Sagan

-------------------------