LS 107 - Beyond Verbal with Yuval Mor

This show is produced by the Hartman Media Company.

For more information and links to all our great podcasts, visit hartmanmedia.com.

Welcome to the Longevity Show. Informing listeners on the important aspects of health, wealth and happiness to ensure you live life to its fullest. With fascinating interviews with top authors and gurus in the field. Along with the latest news in the science and technology of longevity, we're going to reveal expert advice and amazing secrets of living a longer.

And now, here's your host, Jason Hartman.

Welcome to the Longevity and Biohacking Show. This is your host, Fernando Aris, in partnership with the Hartman Media Company, filling in for Jason Hartman Today. Today it is my pleasure to welcome Yuval Moore. Yuval is the CEO of Beyond Verbal. Beyond Verbal is a company that analyzes emotions or from vocal intonations. This is very cool and has many practical applications. Uval has 25 years of experience as a senior leader in early stage high tech companies, leading them to successful exits. Yuval is also a member of the Hive Steering Committee. The Hive Steering Committee is a non for profit accelerator which which helps immigrants and returning citizens start new companies in Israel. So welcome Yuval. It is a pleasure to have you on the show.

Thanks a lot, Fernando. And thanks a lot for the opportunity to speak with you and with the audience. I'm actually very thrilled to have this conversation, so thanks for having me.

You're welcome. I would like to give the audience a sense of geography. Where are you located today?

So I'm actually based out of Israel, so right now I am in Israel in Herzliya, which is the town right north of Tel Aviv. It's actually pretty late my time, but because we're doing a lot of business with the U.S. these are our regular working hours.

Yes, I know Herzliya very well. I've been there a few times on business trips. It's quite an affluent city in the coast and I can tell the audience that it has a robust startup and an entrepreneurial type culture. I've stayed at the Den Acadia Hotel several times right on the beach. It's just a beautiful place for jogging, swimming, water sports. And of course, as I mentioned, the culture there is just amazing. So I'm very happy to have you in the show and talk about your company.

And just to say one more word about Herzliya, as you alluded to it, if you walk the streets of Herzliya as you have, in some way feels like Silicon Valley, because you see the offices of Apple and Microsoft and Salesforce and a lot of other High tech companies that are based there as a result of all the high tech activity that are happening in Israel, which is why it's called the Startup Nation. So very, very exciting area to live in, for sure.

I agree 100%. There is no exaggeration whatsoever. It does feel. I've been to many of the buildings that you just mentioned. It's definitely a great culture and just a great environment for startups. Now, let's talk about your company, Beyond Verbal. What struck me as quite interesting is your website has several verbal analysis examples. It includes the late Steve Jobs and our new president in the US Donald Trump, where your product is analyzing their speech. And it's quite fascinating. Now explain to us how one can derive specific emotions from voice intonations. How does it actually work?

So if you think about this, us as human beings have been able to analyze emotions for, you know, it's a capability that evolved over millions of years. A lot of it has almost to do with like, survival skills, the ability to understand our other people, to express how we're, how we are feeling. Our research started more than 20 years ago by our chief scientist, Dr. Levanone, and what we did with Beyond Verbal, that started only four years ago is we've acquired the technology that Dr. Levanone developed and put more energy and more modern tools into it. And, you know, going back to what I said earlier, we as human beings have the ability to understand these emotions. And basically what we are doing now and what the new tools, the new technologies like machine learning and artificial intelligence is giving us the capabilities to really model the understanding that people have about emotions to a computer and software capabilities. And in a sense, that's what we are doing. We have a team of psychologists that are analyzing and listening to the different voice samples that we have collected. We have over two and a half million voice samples in our database, and they label them and tag them. And then we are using all kinds of unique capabilities that we've developed over the 20 years in order to model the understanding of the emotions into software, into an engine that we've developed.

You mentioned survival skills. And I've always been fascinated by the fact that dogs, for example, they don't quite seem to listen to the words themselves. And I speak several languages and I've noticed this. They seem to listen to the intonation that is given to the command or to the voice. And this is innate in us, isn't it? It's something that we evolved with. So if you can tap into that seems like a natural way to get communication going. And I think that that's what you were alluding to, correct?

That's absolutely what I'm alluding to. It's interesting because usually I try to steer away, or at least not so early in the discussions about the capabilities of dogs. But what you're saying is absol. Correct. And now we can talk about this because there's been a lot of research going on also about the ability of dogs to understand emotions or emotions as they are being projected by the tone of voice. And interestingly enough, it is being done in a very similar part of the brain that is the area, the center where human beings are analyzing emotions. And this is the limbic system. So it's not part of the cognitive part of the brain, but it's the precognition, if you will. That's the limbic system. And this is where instinctive emotions are being generated and being analyzed. So in a way, it is very similar to the way dogs are understanding the tonation of their owners.

Now, I imagine that nowadays with modern tools, both hardware and software, it allows this sort of technology to be more accessible. I imagine that this could not be done a while ago without huge amounts of hardware. And even more so, it couldn't be done real time. I can see where even mobile platforms, as we go in more depth with the system, will be available at a mass scale. So that brings me to a question. Do you do real time analysis and is this already done by the device that the person is using or is it done by some background computing engineering?

So you're absolutely right that we can really benefit from the fact that technology is playing in our favor all over. So the proliferation of mobile devices, the quality of the microphones on the mobile devices, the bandwidth that networks, the cellular networks are now providing to the mobile devices, and the fact that now there is this whole platform of cloud computing, all these capabilities basically enable us to take any device that has a microphone and Internet connection and turn it into an emotions and we'll talk about also the other capabilities but emotions and wellness and even health sensor. In our case, we are doing the analysis in the cloud. It can be done on the device itself, but from the ability to really scale and to be able to add new capabilities. Everyone is now using the cloud because then every time we have a new algorithm and improvement to the engine, we can immediately deploy to the cloud and our customers can benefit from it very seamlessly. So it is going to the cloud, the analysis is being done over there and then the results are being sent again to the mobile device or any front end device that is using it.

It's amazing. You've mentioned earlier that the limbic system is where our instinctive emotions are placed. And what is fascinating to me is that there is a study and there's just an amazing area that your company is involved with between voice analysis and medical conditions. Can you tell us a little bit of what's going on in this amazing area?

We are actually now taking a lot of things that in a way come natural to again us as human beings. And we'll talk in a second about how we are stretching it beyond that. But if you think about it, we come across a lot of cases where we tell our friends, our family members, hey, you don't sound so well. What's going on? And this is usually our instinctive ability to tell that something is different. And the person is maybe, you know, very tired or having a flu or, you know, is not feeling well in general. Maybe they went through an injury or something like this. But then if you take it to the next level, then we can definitely think about condition, medical conditions that are coming, like neurological conditions, if you will, things like Parkinson or ALS or autism. All these conditions, and I'm not even talking about mental health conditions, these are things that with the human ear, we can listen and we can tell that the person that is speaking.

Sounds different. And really what we are trying to do with these conditions is to be able to identify the specific characteristics that makes the analysis specific enough to be able to tell this person maybe has Parkinson's, this person maybe had dyslexia, this person has ALS and so on. And these are all neurological conditions that influence the vocal cords, and as a result, they manifest themselves in the voice of the speaker and you can really hear the differences. So now we can take the same basic capabilities of analyzing the voice and extend these capabilities not just to analyze the motion, but to also being able to understand medical conditions.

Now, I know, you know, this is innate to us human beings. You know, we can hear people that are not sounding so well. And I've seen this firsthand. I think most of us or all of us have seen this. When people don't sound like themselves, you can just. There's a sixth sense that kicks in and you can tell that there is an issue going on. And what's amazing to me is how can you break that down and somehow code it in software to be able to recognize the subtle changes or, you know, what are the triggers that will indicate that there is an issue? I know that this probably can get very technical. But what do you look at or how do you break this down and code this program to tell you that there is an issue going on like a human would.

So first of all, what I want to like this whole thing that you said that this is innate to us, which is totally true. But I want to also take you to another interesting domain, and this is Chinese medicine. All Chinese medicine, when people are going or used to go to a Chinese doctor, the doctors there would be using some fascinating methods to really identify medical conditions with their patients. They would start by looking at the patient in general, you know, saying any special signs. They would also do things like smell the patient. And now we actually know that there are different areas that, you know, dogs could send some specific medical conditions that people just by. By smelling and how people are trying to model this into products as well. And the other thing that the Chinese doctor would do would listen to the person, the patients that would come in, but again, not to what they say, the. But how they sound. So there is definitely now, we are now closing the loop of going back to even some very old and well tested methods that doctors have been using in the past. Now, if you take this on the intuitive level, then you can definitely see how it should be possible to model these specific voice features that, you know, these old doctors would listen to, but in a much more accurate and in a much more subtle way to really get to some specific patterns in the voice. And really now the challenge is to make sure that the analysis is specific, as I mentioned, to be able to differentiate between one condition to another. And if the voice, the condition is manifest itself in the voice in different ways for Parkinson's, then for autism, then a computer software with machine learning and so on should be able to really model this in a very accurate way. And that's basically what we are doing.

It's always amusing to me to see this connection between old medicine and the knowledge that exists, as you mentioned, in Chinese medicine, where the smell or the voice, the intonation, I know that they look at your tongue to see what sort of coloration it shows. It gives you these clues that form a picture. And with the technology as it evolves, it allows us to encode that information into a machine. And the machine has a lot of resources for pattern matching that will give us this amazing ability to do partially. And I'm sure that in the future we'll do a faster job of recognizing some of these issues with patients. Now I know that the Mayo Clinic has also been involved in some research and there's been some recent results that look at voice symptoms and heart conditions. And I know you're involved with it. Can you explain to us what that research is about?

Absolutely, because this is the really fascinating area that is extending yet again the limits and even the imagination of some of the things that can be done with analyzing the voice. And by the way, the reasons we like the voice so much is because it's readily available. And this is the tool that you can use on an ongoing, continuous basis, but also in a non intrusive way, so you don't have to attach any device, you don't need to touch the customer. It's completely non invasive. And the voice is actually one of these very expressive tools that in a way have been overlooked for many years from a technology perspective. Now, specifically what we did with Mayo Clinic is based on some preliminary results of testing that we've done in Israel with some hospitals. We basically engaged with them over the last two years with the research that the objective was to see if we can find specific voice features that can correlate or can distinguish between people that are suffering from coronary artery disease, cad, and people that do not. So what Mayo Clinic did is they've signed up about one hundred and fifty patients for this research. 120 of them were referred to the Mayo Clinic to do an angiogram. And then they had two control groups. One control group was of healthy people, healthy patients, and the other control group, which is very interesting, they chose also people that came to the Mayo Clinic to do another type of operation, but not related to the heart. And this is in order to see that the features that we find are specific for heart condition and not just because the people are stressed because of an operation. The bottom line of the study is that they were able to find, Mayo Clinic was able to identify. This was a double blind test. So we didn't know the results. We don't know anything about the patients even to this date. They were able to find some feature that ended up being very indicative, very specific to be able to say if these patients had a risk for suffering from coronary artery disease. And this is fascinating because this is the first time that this is being done with conditions that are not neurological conditions, these are non neurological conditions, this is a heart problem. And this is really expanding the imagination, if you will, of what you can do with this type of technology. So the results, you know, they are preliminary. We still need to do, you know, testing with other languages and so on, which we can talk about. But the first such study, very fascinating and promising results.

That is amazing. That is like you said, it's non intrusive. It's so easy to do. You know, you're not involved in drugs, there's no side effects. It's just a very clean way of diagnosis. I just commend beyond verbal for your technology for helping essentially cure heart disease. This is just a start, but it's very promising. That is unbelievable. But one area that I want to ask you is what's the reaction in the marketplace? How, how revolutionary is this new method for diagnosing early stages of progressive diseases? How is it being received in the marketplace?

So as you said, this is very new. And that's why from our perspective, from beyond verbal perspective, we actually haven't talked about this. Even though it's something that we've been dealing with for eight years now, we haven't talked about it publicly until we could really substantiate some of the findings with such a reputable partner as Mayo Clinic. Because A, you get a lot of skepticism with this and B, you really want to make sure that you are heading in the, in the right direction and you do it step by step. I think that the potential that this capability can have on things like remote home care or remote health care, really giving people with chronic disease or non chronic disease to be able to, to do monitoring without the need in the future to always go and see a doctor in person. One of the biggest issues with conditions like this is the need for compliance. And here there is no compliance issues because this, like we said, this can be done in the, in the background on an ongoing basis. So I think it can be really revolutionary and we just need to take it one step at a time and really substantiate every step in order to get the cooperation of doctors, hospitals and so on, actually have to commend Mayo Clinic for this, for the, you know, even the ability to embark on such groundbreaking research.

Yeah, it is refreshing to see such a respected institution be open to essentially what we know intuitively and put that into practice and start making a big difference again without being intrusive. It's just a simple way, one of the most straightforward ways to do diagnosis. You know, if doctors are listening to this right now, what sort of vocal biomarkers should these doctors start paying attention to?

So this is where this whole area is becoming really interesting. Because when I was talking with people about the neurological conditions, it was one thing, but when we were talking about the non neurological like the heart condition, then in a lot of Cases when people were asking us, so what really is this vocal biomarker? It's not really something that is easy to explain. And the reason is unlike neurological conditions I mentioned Parkinson, als, dyslexia and so on, where you can hear what the distortion in the voice is when it comes to the heart condition, this is not something that you can hear with a regular human ear. And the example that I gave, in much the same way that the human eyes can see a certain range of frequencies, but we cannot see cellular, luckily we cannot see cellular signals, WI FI signals, ultraviolet, infrared and so on. In much the same way the human ear, it sends the to a specific range of and can understand specific range of frequencies. But some of the things that we are finding now with the heart conditions, it's something that is beyond the range of frequencies that we can understand with the human ear. And so it's not something that we can intuitively explain, but machines, computers and so on can do this type of analysis. So these vocal biomarkers, as you mentioned, and this is the term that we are using, which is a great term, is not something that you can use the regular doctor's ear in order to listen and understand.

Again, it's amusing and amazing to see technology coming full circle and being able to extend our ability, our human ability to in this case diagnose heart conditions. And of course with the emotion analysis to be able to do this in a real time fashion, which has not been able to be done in the past. Now for the voice samples, I know that online you do have ways for people to try out your emotion analytics. So when you mentioned the millions of voice samples that you've received, are most of them coming from people online that can just go to your website and try out a voice amp and see what sort of emotions come out.

So actually the most common use that we are using right now for demonstration capability and evaluation capability is not using our website, but using an application that we have more like a business development or demo application that is called MoodyV M double O D I E S And this application is available for free on Android and iPhone. And this, we are using this application mostly, as I mentioned, as a business development tool. People that really don't believe that it's even possible to do, we tell them, hey, download the application, give it a try and once you've seen what is possible, because this application is, is only showing a very small portion of our capability, get in touch with us and let's see how we can do business together. But in the application we give people the ability to vote on their analysis that they see as a result if they agree or disagree or don't know. So this is one source of voices that we get and then we can use it from a variety of use cases that are currently being tested or deployed. Things like a market research or call center or smart home and so on. So we do get voices from a diverse number of sources and we can also use just things that are available on the Internet. YouTube is a great resource and so on. So once you have the mechanism and process in place, it's fairly easy to to get really variety of sources for voice in different languages and so on.

That is excellent. Yeah, if you just search for beyond verbal in the App Store or in the Google Store, you'll be able to get the mobile application. Which brings me to an interesting question. You know, Apple has Siri as an AI artificial intelligence assistant and I can envision where Siri would be able to take advantage of of emotion analytics to help determine how you are and interact with a person at a deeper level. Are you seeing work being done there? Are you involved with any of the work for AI?

So I think that the reason we are so excited by the opportunity that exists today is because of what you are saying right now. We spoke before about the technology enablement part, but now if we're gonna move to like the market demand or the development that we see in the market, we see more and more devices that have primarily a voice interface. Many know Siri or Google now is one. One example, humanoid robots, smart homes, cars, you know, it's almost a carded, almost like a hormone, if you will.

Yeah, the ultimate mobile platform.

Exactly, exactly. So we see more and more use cases where the voice is the primary mode of communication. You know, people ask us, but now people are moving more and more to text with, you know, WeChat in China and WhatsApp in the rest of the world. And actually what these companies are saying, you know, also with the Facebook Instant messenger is all these platforms are adding voice message capabilities because people really like to use voice as a means of communication. So the more voices we see out there, we definitely see the need to add an emotions understanding on top of just understanding the words. So the way I kind of describe it, a few years from now, I don't know if it's two years, three years, whatever, we probably wouldn't call our phone smartphones if they wouldn't have. On top of understanding of the words, also understanding of emotions, it has to be to find its way to really all these use cases. And that's why we really believe that it's just the beginning. We're just scratching the surface of what is possible.

I agree 100%. Being in the technology field myself. It's as if if we are missing an entire dimension when we talk about intonation or the lack of ability currently from the AI devices to get what the person's intention is. Once you reach that ability, it opens up this amazing possibility where the intonation can determine what the voice is intention.

Was just wanted to say that I think this would open up the possibilities. I mean, just think about circles of friends or families. Imagine that you are, you know, you are based now in Arizona and you know, your mother is on the east coast and imagine if she's part of your circle and you can get some kind of notification that says, hey, your mother just, you know, she doesn't sound so happy today. Why don't you give her a call, send her flowers. Or, you know, imagine if you're on the way home and you get a notification that your wife is really had the rough day today. And it will give you, you know, the options of go buy her some flowers or maybe skip this all together, go have beer and just come back much later. And so yeah, yeah, it's social interaction can get a whole new dimension once you add also the emotions. Understanding part of it.

Amazing is incredible. I know we're coming up to the end of the show and it's just been fascinating, you know, this whole topic, I think it's about to explode, certainly on the medical diagnosis piece and on the social piece and I can see exactly how that could play out. So before we end the show, could you give us audience any URL or any links where people can learn more about your company and the technology behind it.

So definitely a good place to start would be on our website, you know, www.BeyondVerbal.com There is quite a lot of resources there. And also link to our YouTube channel in our Twitter account and so on. So that's definitely the best way to start. And then again, play a little bit with the application. And now we've recently added, as I mentioned before, we didn't talk about this earlier, but now we've also added information on the health side and work that we've done not just with Mayo Clinic, but with other institutes in the past. I guess as my last kind of a comment, what I wanted to say is that the vision that we have is that we would see applications that are using our technology as kind of in a way like a personal guardian angel, if you will, and something that you know on a daily basis or while you are having your regular conversations or like we said, driving in the car, it can give you indications of almost like a red, yellow, green, you know, green, it looks like everything is okay, you can go about your day yellow with something different may be worth checking in red. And this is really the difference where we do not see this as necessarily, at least not initially the diagnostic device, but more like as a long term monitoring device. The red would actually mean, listen, we've identified something that is significantly different and we can maybe even assume associated with a specific condition. Why don't you go call your doctor? Because I think that once you go to the doctor, they have very good, probably better diagnostic device. And this should be more of like a screening and early indication type of capability. So we would all have our own personal guardian angel and it will give us the indications of how we are doing.

That is a very good point and I can see the value of that and it's been a fascinating topic and I again, thank you, Yuval. It's been a pleasure talking to you and you have a great day. Okay?

Thank you very much for the opportunity. Enjoyed it a lot. And good evening to you.

Thank you so much for listening. Please be sure to subscribe so that you don't miss any episodes. Be sure to check out the show's specific website and our general website, hartmanmedia.com for appropriate disclaimers and terms of service. Remember that guest opinions are their own and if you require specific legal or tax advice or advice in any other specialized area, please consult an appropriate professional. And we also very much appreciate you reviewing the show. Please go to itunes or Stitcher Radio or whatever platform you're using and write.

A review for the show.

We would very much appreciate that and be sure to make it official and subscribe so you do not miss any episodes. We look forward to seeing you on the next episode.