Dr. Achin Bhowmik Details the Many Ways Today’s Hearing Aids are Not Like Your Father’s

Our Chief Technology Officer and EVP of Engineering, Achin Bhowmik, Ph.D., tells Dave what brought him to Starkey, how AI is a hearing technology game changer, and the many ways his team is turning hearing aids into want-to-have devices.

Link to full transcript

Transcript

SPEAKER_00 0:02

Welcome to Starkey Soundbites. I'm your host, Dave Fabry, Starkey's Chief Innovation Officer. Our guest today is Dr. Achen Bulmick, who serves as Chief Technology Officer and Executive Vice President of Engineering here at Starkey. Of course, he's uh become a good friend over the last several years and colleague. Uh, thanks for being with us here today, Achen.

SPEAKER_01 0:22

Thank you for having me, Dave. And it's a fun pleasure to talk to you no matter what the stage is.

SPEAKER_00 0:26

Indeed. And we've been on more than a few stages together. I was thinking about this earlier, and I think we've been on at least three continents together as we've launched uh Livio AI when I first rejoined the company and you were here. Uh and it was really the industry's first device that incorporated embedded sensors in combination with artificial intelligence, and it's been a rocket ship ride ever since then.

SPEAKER_01 0:49

What a journey it has been, isn't it?

SPEAKER_00 0:50

It sure has.

SPEAKER_01 0:51

I I can't wait to get back on normal time so we can hit the road again.

SPEAKER_00 0:55

I do miss that. As much as I love these podcasts and I love the virtual trainings that we've been doing and launches that we've been doing, uh it it there's nothing quite like having the engagement of an audience in front of you to feel their excitement as they're responding to this uh technology that we're launching. Trevor Burrus, Jr.

SPEAKER_01 1:11

Part of our job is to connect with the customers and patients and audience around the world so we can test what the user requirements are and how the products can fill more of their needs than they have done in the past. So that's a part that uh we can do a better job as as the world gets back to normal.

SPEAKER_00 1:29

Yep, fingers crossed that we'll be um somewhere near uh our audience members soon. And uh let's talk a little bit. We'll get to the technology because I know that's both of our favorite topic. But uh I want to talk a little bit about your background in terms of how did you find your way to Starkey. Right. Um and uh you know, maybe begin with the more recent, uh, your more recent past at Intel. Um how did you how did you find your way to Starkey?

SPEAKER_01 1:54

Oh, that's a good question. So I think uh the way that I would I would uh start is by transcending myself of the companies and the structures and the processes, but um ask I ask myself of what do I find to be my calling in terms of all of the learnings and passion, where do I want to make the mark and an impact that uh you know they all has been said about you have to find your passion such that work doesn't feel like work. So for me, I um I'll I'll introduce some jargons, I'll define them. Um uh I call that perceptual computing. First, I'm an engineer at heart. I'm not a neuroscientist, but spend a lot of time with them. And early part of that uh started with this desire to uh get into natural human-computer interaction. And I mean what's natural? Essentially, how do humans do it? How do we do it in our normal lives? When you start to ask questions and peel the layers and get deeper into it, there is no escape but to uh get an understanding of how our systems work. How does the human system work? More specifically, how do we sense and perceive and understand the world, which then leads us to interacting in the world the way that we do. And uh as an engineer, you look at all the components that we have. So the the term that I have embraced along with other colleagues in the industry back then and at Intel when I started this initiative called uh perceptual computing, uh the whole idea there was to mimic biology. And a way that I explain to the general audience is that uh you know, we we we take our perceptions for granted. It's just it comes so naturally and so easily to us. We wake up in the morning, open our eyes, and a flash of light comes in, but we totally are not uh aware of all of the physical processes that go in. When you consider the physics behind it, the light, the rays or photons that essentially come from the sun and you know, uh through the light in the ceiling that get bounced around in the environment around you and enters our eyeball, which get transformed at the back of the eye in the retina uh to into electrical signals in the form of neural impulses, which then find their way into the into the visual cortex first and then gets interpreted there and gets into a lot more sophisticated processing for us to to understand what's going on around us. Combine that with other senses.

SPEAKER_00 4:35

Yeah, I mean that's just one sense that's the one sense for vision. Separating that out for vision, and then you think about the ear. I mean, we've we've been saying, you know, you you don't hear with your ears. Your ears act as sensors as your eyes do. And then we have that whole process, but then in particular for hearing, vision and hearing are so inextricably intertwined. People say, Oh, I'm I'm losing my hearing a little bit now. I better learn to lip read. Well, we've been lip reading our whole lives and using that information, and they both serve as sensory inputs, the the four the two eyes and two ears.

SPEAKER_01 5:08

The key is the multimodal aspect. Yeah. Because we we are we have all of these sensory modalities on all the time, and not to leave behind touch, smell, taste, and then the higher levels of deriving semantic information from it. You know, Dave smiling, the other person's looking confused. So it's combining all of the modalities of sensor data, eyes, ears, skin, the touch from the skin, the somotric sensation, smell, taste, and then this amazing, our amazing ability to create a model of the world with all of the sensory inputs, and that allows us to navigate in this world, understand it, interact with each other. So that's the whole field I call it perceptual computing. And then the way I look at it, um, the parts of, you know, we all got trained in science in high school and then the college, physics, biology, and all of that. And then when you start to get into practice, you know, it led myself into this natural human computer interaction world where there was a clearly clear, almost like a demarcation of the old world of computing and the new world of computing. And I'm just uh playing back to you the strategic thinking that went on in my head, along with esteemed colleagues and mentors at Intel and other larger companies, Apple, Google, and mentors in the Silicon Valley area. Uh, to us it was clear that the world of computing, as it evolves, and there's not a simple line that divides the past from the future, but there was this clear sense that the world of computing was fast evolving from systems that uh are uh stationary on your table, uh, and then it became the computer in your pocket. And then the computers are going to be just everywhere. The concept of pervasive computing, where you know the distributed uh sensing, computing, interaction, think of cars that drive by themselves. What is that? It's a computer. Drones that can fly around without running into objects because it recognizes what is around it. That's a computer. Hearing devices that are in your ear, that's understanding what's going on around you, the acoustic environments, that's a computer.

SPEAKER_03 7:14

Right.

SPEAKER_01 7:14

So the definition of computer was evolving. And to me, these were all perceptual computing in the way that just computing and logic was not good enough. You needed the computers to develop senses to understand what's going on around you. So that I think was the bridge between when I consider that to be the bridge between my old world in my career at Intel to the new world at Starkey. I don't see the difference between the two.

SPEAKER_00 7:37

Right.

SPEAKER_01 7:37

The fundamentals are the same.

SPEAKER_00 7:38

Trevor Burrus, Jr. Yeah. Yeah. And and you say, you know, the the bridge between old world computing, new world computing, and I would even argue that the bridge between machine and human, there's no line anymore. I mean, I'm a good example of that. I'm slowly becoming a cyborg. Suddenly we joke about that a lot. You're a living cyborg. Yeah. And this issue of now your evolution, if you will, into uh our company in our world in the hearing space, uh. But knowing, again, that there's integration from sight and even touch, I'm old enough to remember hearing devices in the early days that use vibrotactile sensation. So right now we talk about vision and hearing, but there was vibrotactile sensors that would be used to help people differentiate between different sounds if they had profound hearing loss. So now the world that we live in has really led to this journey that since you joined us now in uh in when we launched Livio AI in 2018, we were the first manufacturer to introduce a product, a hearing product with embedded sensors in combination with machine learning and artificial intelligence, and we've been off to the races.

SPEAKER_01 8:45

And I'll I'll uh tell a little bit of the story from behind the scenes for your audience. Yeah. Um you have heard me uh say these things many times, but for the benefit of your audience, I'll just uh say uh a few bit more details than I've shared in the past. So suddenly when I got to interact with Mr. Austin, uh I remember the frank conversations we were having, where he would say, I don't know much about this thing you call perceptual computing, but I like that. It's like uh learning from biology to influence technology. But then he said that with that, I I see what you're doing is you are using that knowledge and intuition to develop technology to make better technology.

SPEAKER_00 9:25

Right.

SPEAKER_01 9:25

So a a car that gets you from point A to point B, but with this sophisticated technology, the car will be aware of its environment and take you from point A to point B. Or the robots, drones, computers that will let you just stare at it and recognize your face. So those are your enhancing technology with the knowledge of, you know, technology that's inspired by biology. How about we make it full circle? His challenge was can we can we then, rather than working on on machines, help humans better lives with this knowledge that comes full circle? Learn from biology, get inspired by how the natural world works, develop technology that's inspired by it, and then use the technology to enhance human perception. So, in ways, the to me it appeared to me to be the opportunity for making a shift from enhancing machine perception to augmenting human perception. And the other ways in practical aspects, as I learned about hearing aids from him and other esteemed colleagues around me at the time, it felt like there was a time, an opportunity for us to uh migrate the hearing devices, which were already playing an amazing role in connecting people with people, um, helping people hear when they can't without the devices. But then we had the opportunity to transform these devices from a single function product to a multi-purpose device that served many more purposes and help people in many more ways than their single function nature. And you heard me give examples of how you know uh uh easy to look back and see an example in the past. Where uh when Apple introduced iPhone back in 2007, they took a device that was already popular, already very useful, the mobile phone. Yeah, but iPhone not only did everything a mobile phone could do, but over time it became your camera, it became your GPS device, it became your pocket computer, it became your YouTube watching device, the media platform, gaming device. For us, there was a unique opportunity because nobody had done it before. And we had even better opportunity to leverage the hearing aids' ability to be with you without being without being aware of it. People who use hearing aids, they it's a killer application. I always people who have to use hearing aids, they will use the devices in their ear all the time. And while it is there, the three aspects. How can you make it do a better job than it has ever done with new technology? Hopefully you get to talk about those sensors, machine learning. And then second, if it's in my ear already, and ear is the best place for sensing health and doing biometrics, why don't we do those and keep me safe? Hopefully it will be aware of my health problems before even I know, or I happen to accidentally see a doctor. And the third, how do we make this device being be the conduit to the world of information and become my very personal virtual assistant? So all of those opportunities are there. So with that, with the the the bridge between the old world and the new world of computers that are stationary to computers that are with you all the time, they're pervasive, you're not even aware of them. And those computers, you know, rather than just being a gadget that helps you entertain yourself, something that help helps you live better lives. So that that is how, in a short while, I would say my journey from the recent journey from the days at Intel as uh as a leader for a new group there working on perceptual computing to the opportunity at Stucky working with you and others in transforming hearing aids into the next generation wireable device that you want to have, not just because you need to have them.

SPEAKER_00 13:11

Exactly. I mean the stigma surrounding uh hearing aids is very different than what you had in terms of working at Intel with drones and computers. Uh people want those. Uh you know, was that an initially was that a challenge when you came to the hearing aid world, given that, you know, no matter where you are in the world, even if you're providing them at no cost to the end user, less than half the people choose to use them. And I think that's a very important point that you raise. That right now people feel like they have to wear the devices, but they're in at least a an approach avoidance, if not an avoidance-avoidance conflict. They're not crazy about having hearing loss or about using hearing aids, but the mission is to transform them into a uh a want-to-have kind of device like you've been working with in the past by increasing that multi-purpose, multi-functionality. In effect, you know, we're we're sort of augmenting, giving them a superpower uh uh not only in the hearing space, but then with the ability to sense health and wellness, uh to be able to use it for as as Jarvis uh as an additional communicator robot uh and in accessing the internet. And what would you say when you joined our company?

SPEAKER_01 14:24

So um uh that's so let me let me start with the stigma side.

SPEAKER_00 14:26

Yeah.

SPEAKER_01 14:27

Because I like I want to add my bit to it.

SPEAKER_03 14:28

Yeah.

SPEAKER_01 14:29

You know, which engineer doesn't want a good challenge, right? So it's all about problem solving. If you think about the world of technology and engineering, there is curiosity part of it where we just want to know things, we want to know how things work. And the moment we wanted to call ourselves engineers, we we take on an additional responsibility of solving problems. And when you explain, uh you know, those I heard a lot in 2017 when I joined, there is this big stigma about hearing aids. I think we still have them, we haven't solved them quite, but what a challenge, right? Because uh if if our device has become so useful and essential for someone who cannot live their lives without it, I have seen plenty of them. You know, people come in struggling to hear, they get a good pair of hearing aids perfectly fitted with a capable professional, and then they leave in tears. Yeah, they cry, right? So suddenly the product is awesome, or the technology connects people with people that they were not being able to do before, and it leaves them in tears. So the value is clear. Well, then there is a stigma. What an awesome problem to work on, right? So how do you go about that? Well, first of all, make cool devices. You know, this thing said about how uh uh when you pick up an iPhone or or a good well-designed piece of uh technology, get into your Tesla car, right, or your Apple Watch, or any any good device, it's got a coolness with it. It's the way the industrial design goes. Yeah, right. And you know, you look at um uh modern hearing aids that we have built now, you've just we are startershipping evolve AI hearing aids. So I'll tell you this, right? When I uh you know, I've told this story long times, many, many times, that I don't have hearing loss, but uh I'm hooked to my hearing aids. And people might say, Oh, well, of course you are CTO for Starkey, you're going to use your hearing aids. But genuinely, there is this value of about this device. So, first of all, when I pull out my hearing aids, and I do that uh in you know in meetings, let's say a conference in Silicon Valley or you know, at a meeting at Apple or something, first of all, people the jaw drop, it's like you had something in your ear. We didn't see it.

SPEAKER_00 16:41

We didn't even see it. Yeah.

SPEAKER_01 16:42

And then I'll show it to them. And they'll say, wow, that's a well-designed device. And then you would say, I can see you wearing it all day because you can forget about a device like this. It doesn't, it has nothing dangling from your ear, it doesn't create a tug, there is no fatigue about it. So the amount of design work that has gone in in making these devices cool over the last few years versus the ugly pictures that you see in Wall Street Journal, New York Times, when they talk about hearing aids, these big devices for the colour.

SPEAKER_00 17:13

They may as well be an ear trumpet. 10 years or 20 years ago, right? Yeah.

SPEAKER_01 17:16

So that's so the devices that become cool. One way to remove stigma, because you know, you see people in the planes now with devices coming out of their ear, and if you have such cool devices, you call hearing aids, it removes the stigma because it's cool devices. Number two, who doesn't like to be uh to be told of a problem that might they might have in the future? That's the path we are on with sensors built in, with machine learning that tracks my physical activities, so I don't need to get you know other uh peripheral devices around my body because these devices are multifunctional and that does it all. And how cool is it, you know, my my gardener who's uh a Spanish-speaking individual, I don't speak Spanish, he doesn't speak much English. But I can double tap and bring up a translator or go to my mobile app and bring up the translator. And my hearing aid, now being connected to the cloud via the smartphone, has an infinite amount of computing resources available to translate from any language to any language, between 27 languages. It allows me to have a one-on-one conversation. I do have to be patient with the 50 milliseconds delay that's required for the cloud. But guess what? Much better than having a human translator in between me and my gardener, right? So this device, I believe we are way on its uh on our journey to make it so essential and cool that the stigma will cheaper way at we'll stick cheaper way at the stigma.

SPEAKER_00 18:38

Well, two things that I'll uh that come to mind when you when you recount that. Um the first time, one of the very first times we traveled together, we went to TechCrunch in San Francisco. And um we they had set up some interviews with various media agencies. And remember there was a TV reporter who came and he was like, oh, yeah. Uh you know, and there's all kinds of cool tech that's around there. And he was like, he was just gonna do a vignette of all different technologies that they were seeing at this tech crunch. And suddenly, you know, he said, Well, you know, what what do you got in terms of hearing aid? And we started talking about the features just as you have now. Suddenly he said, Really? You can do that. And then we had a remote mic and we were remembering we were playing. He did a whole segment on local TV that night. I remember because we nerdjacked the conversation.

SPEAKER_01 19:23

This is a hearing it company. Let's tell me what it is because I need to go check so many other things. And he ended up spending an entire evening with us, and then he packed up and went home and did the entire episode on us.

SPEAKER_00 19:32

The whole segment was on us.

SPEAKER_01 19:33

So that that's our you know, the secret. Yeah, we need to get it out there, let people know of this amazing technology that enhances their lives. Which is the same.

SPEAKER_00 19:43

The litmus test for me too, on a plane, sitting next to Millennial while I'm finishing up a phone conversation with a colleague or my wife, and then they say, Were you just using those hearing aids for that? And then I say, Oh, but just I'm just getting started. And when someone says, I don't have a hearing loss, how do I get that? Yeah. It's sort of that's the battle that we've been facing is fighting the stigma, improving first and foremost the people's ability to hear better, but then translate that into live better through uh providing these health and wellness features and the communication features. Well, let's transition into the technology a little bit. You come from this world where AI is commonplace, but to the hearing aid space, I mean, can you talk a little bit? I mean, I say on the one hand, in our industry, AI is ubiquitous and yet it's meaningless. Everyone's using a little bit different definition, a little bit way of looking at how they define AI, machine learning, deep neural networks, or deep learning. Can you talk uh uh in simple terms about the differences between those and how we're defining it?

SPEAKER_01 20:45

I think all I need to do is like explain the paper you and I wrote recently at uh Seminars of Hearing and also IEEE Computer. So that would I would I would approach it for your audience from the viewpoint of what is in there for me, the user. So there is the world where everything needed to be set for the uh environment that you are in. So now let's discuss that a bit. So let's say um I have my hearing devices and uh it's designed to take in sound and amplify it. Well, it sounds very simple, except in real world, the sound can mean many things. We know in our ear, the conduit to the mechanical vibrations of air, and it's getting a what I call a cacophony of sounds from all different sources. You know, fortunately you and I are having a quiet uh having a conversation in a quiet place now. But most often you found uh find ourselves in a coffee shop in the restaurant where uh it switches this the nature of sound switches from being sound of interest to sound of source of annoyance.

SPEAKER_03 21:53

Right.

SPEAKER_01 21:53

We are having a conversation right now. I don't care so much about the music around us, but we'll soon be Taking a pause and then I want to enjoy that music. And the device, if it was dumb, if it was amplifying all sound that's that's coming into the ear the same way, it would be an annoying device. You know, when we'll be talking, it will amplify your voice, and then there is uh background music, it will be amplified as well, and make it will make it really hard for you to understand what you're saying. Imagine somebody suddenly bringing a vacuum cleaner and starting to clean a track passing by, or as soon as I step outside, there's a big wind whoosh coming in and uh creating sound that will get amplified. That's the world of the old hearing aids. What AI can do. And when I explain what it can do, people say, Can I have it now? So we get often caught in technology, jargons, machine learning, deep learning, artificial intelligence that gets scary for our uh patients. But if we explain to them this way, imagine a device that is smart, intelligent. It knows what sound should be amplified, should be of interest to you. Because we are now having a discussion. Perhaps the speech component of the sound should be amplified. And then the background music should be decreased or not amplified. And then the moment I stop talking, perhaps I am now paying attention to the surrounding sound. So they should be given a bit of priority. Or, you know, I am now stepped outside and it should automatically detect the presence of wind noise and suppress it. That's what is possible with machine learning that was not possible beforehand. Why is that? Because the old world of computing prior to AI was all about engineers coding up the rules, saying, you know, it's this if this then that. Right. The conditional statements that sort of was the heart of computer programming paradigms prior to machine learning and AI. I needed to code the rules into the program for the device. If you see this, do if you hear this, do that. If the signal has these patterns, do this. And I could do that for maybe a dozen rules. Maybe I can put in a hundred rules, maybe I can put in a thousand rules and have gigantic programs that require a lot of memories in the device. Or we could rely on this really exciting development that's taking place in machine learning. When you pick up newspapers or listen to news and you hear AI is taking over the world. It's been called the new electricity. AI is the new electricity, Andrew Ying at Stanford, a good colleague of mine, said. And um there's truth in that. I I, you know, for a longer time I would get into more detail, but believe me when I say the walls with the world, the way it's shaping up in every aspect of our technology is going to change with AI. Cars are going to drive by themselves and fatalities are going to be orders of magnitude lower. And you're just look at the junk emails you used to get three years ago and versus today. Most of the junk emails get automatically understood to be junk and categorized as such. Without that, you're you'll not get the Gmail experience that you get today. Or keeping your bank accounts safe, because AI can do much better job in detecting anomalous transactions and is doing a better job than human doctors in reading radiograms detecting cancer. In our own world, what does that mean? The devices with machine learning algorithms can do millions of automatic classifications. And you know, we we said these numbers and people got stunned.

SPEAKER_00 25:36

Yeah.

SPEAKER_01 25:36

The device makes 55 million automatic adjustments every hour.

SPEAKER_00 25:40

This is an Evolve AI. Evolve AI product. 55 million.

SPEAKER_01 25:43

Yeah, for me, it's a billion adjustments every day.

SPEAKER_00 25:45

Right.

SPEAKER_01 25:45

Think about if I had to make a billion, forget a billion, a million, forget a million. If you have to make a hundred different adjustment adjustments every day, impossible. I just want to go about my life. I want to talk to you, I want to say hello to people. I want to be in a meeting. I want the device to be in the background, be so smart that it makes a billion automatic adjustments based on its artificial intelligence. And that's what our devices do. So when when people talk about AI, it's not 10 years from now. It's now. It's what we're making it thanks to our work the last two years, three years. Right. All and that's just imagine, just close your eyes and imagine what we are going to do two years from now.

SPEAKER_00 26:19

Well, and and think about that, like you said, um continuous improvement. Uh we have very highly accurate automated classifiers now. So that just for people that have worked with hearing aids for quite a while, like I have, uh it used to be that if we had uh directional microphones or noise management, people had to access a separate program when they went into noise and they switched programs. Right then they went if they wanted to listen to music or they wanted to be in quiet, they would switch those programs manually. Now machine learning uh AI does that automatically. But yet the problem that you raise is that sometimes speech is a signal of interest, sometimes it's a noise, sometimes music is a signal that I want to hear, and other times I don't. And I may even want to use different algorithms to process that music so I can enjoy it fully.

SPEAKER_04 27:07

Right.

SPEAKER_00 27:08

The issue is even the best automated classifiers are challenged to uh be perfect given those last two scenarios. The speech as a signal or speech as a noise, or music as a signal, or music as a noise. So talk a little bit about the way that we developed edge mode as a way to really provide this human machine interface that puts AI at the po at the patient's fingertips, if you will.

SPEAKER_01 27:34

Yep, that's a good question. So uh one thing we haven't found out, the entire scientific community yet, yes, there's a lot of excitement about perhaps there is uh some way to peek into your mind and see what you're thinking. I could be sitting here and I I want I want to enjoy music. And uh in other environments, I just want to pay attention to what you're talking about. So there's this intention that I have that only I know. And then you want that AI system to be such that it should take simple cue from me. And so for us, the first instantiation for that uh is edge mode, which is, as you know, is wildly successful. Yes. You know, people are just un unbelievably giddy over it. So you doubt you just simply double tap, and despite the complexities in the acoustic environment, it's going to prioritize speech and clarify things. And it's powered by AI. So we're able to take a uh snapshot, and the system automatically determines what are the best parameter settings for that environment. Because you know the what's fundamentally most exciting about AI is its ability to continuously learn from data, learn from experience. It's like a human child, right? So we learned along the way, right? Um first time I saw perhaps a flame, I might have put my finger in it, and then I you know I realized not to do it again, right? Right.

SPEAKER_00 28:53

Or even more primitive than that, uh a baby um uh doesn't have any rules for how to understand or learn language, whatever, wherever they're they're growing up in the world, but they have their parents um and they see them saying some things and reacting if the child is starting to utter back to them. And the first time they happen to say uh mama or dada and the mom or dad reacts, um they get reinforcement. They don't have any rules, but they know when I did that, she laughed or she approved or she picked me up and carried me around. And that acquisition is truly deep neural network that they don't they don't have any rules when they're starting. Is that the differentiator between machine learning and the technical term for that?

SPEAKER_01 29:35

We would say is learning from data or learning learning from experiences. Yeah. So when you say we learn from experiences, essentially what we're saying is that you know, when you get exposed to more information, we're able to tweak the model of the reality in our head. And so that's ex essentially what a learning machine uh should do. It continuously should adjust, adjust the model with new information and new data.

SPEAKER_00 29:57

Aaron Ross Powell Yeah, and in turn, we continue to improve it. I mean, initially we took advantage of the fact that we have sensors embedded uh in the devices that are worn in or on the ears, and in the same way that they track physical activity, social engagement falls, as you mentioned, they can also pick up a double tap. Um we did have some patients say that rather than always having to double tap, which is not all that obtrusive, but they would also like to be able to discreetly engage edge map uh edge mode in the application. And now in Evolve AI, uh we can do that too. So they can very discreetly just activate that when they're uh uh uh when they have their cell phone with them.

SPEAKER_01 30:35

Aaron Ross Powell, Jr. Yeah. And the world is such that uh the mobile phone has become uh ubiquitous part of many patients' lives. And everybody's familiar with the apps now. They're using app for controlling everything, and we should just respect that they might want to use the app as a control for for things, including invoking edge mode when they need that extra help for understanding speech.

SPEAKER_00 30:54

So I'll ask you the question that we always get asked. If edge mode is so good, why don't you just simply use edge mode all the time?

SPEAKER_01 31:00

That's a great question. So um just think of think about the optimization that the device does for your surroundings. If you're staying in that environment for a long period of time, it would be optimal setting for you. So let's imagine the uh uh situation where I am in my study room and just two or three people around me and talking, and there's this little background sound. And I chose to invoke edge mode for that environment and I double-tapped. The device recognized that as a request to optimize for that environment. And then I walked out and went to a different environment. I could choose for the device's automatic classifications to be good enough for me. But if I need that extra help for that challenging environment that it has not been subjected to before, I might double-tap again and have that device quickly capture the new acoustic environment around it and optimize for that environment. So it's that you call it putting AI in the fingertip of the patient. It's getting amazing feedback saying that I feel like I'm in control.

SPEAKER_00 32:01

Yep.

SPEAKER_01 32:01

When I want to, you know, I need that extra boost for a particular challenging situation. Um my device can do that. It's the AI that listens to me.

SPEAKER_00 32:08

Yeah, and edge mode right now is optimized for audibility. Some people may not always want to be optimizing for audibility, but for other purposes at other times. So and and we continue to uh improve this feature and the flexibility and the automatic aspect of that as we uh develop initial uh initial uh uh applications and future applications.

SPEAKER_01 32:30

That's a good point you bring up. So I I I uh often explain this as the two ends of the spectrum for uh speech. If you prioritize understanding of speech, then we might do signal processing very differently. Right. On the other hand, you might, other side end of the spectrum, you might want to have it sound natural.

SPEAKER_03 32:49

Yeah.

SPEAKER_01 32:50

That might be at the expense of uh understanding speech, like uh uh word error, error rate and stuff. So that the where the dial needs to be should be subjective, individual. And AI needs to be smart in understanding when to prioritize speech, or maybe can get a cue from the user that now I want to understand speech better. Or no, now I am enjoying the ambient sound and I'm actually sitting in a cafeteria, and like you said, the background noise actually is what's my signal at this time. Yes I'm enjoying it. So don't suppress it.

SPEAKER_00 33:22

Yeah. And so don't give away our entire product uh roadmap, uh, but uh stay tuned because there's more to come.

SPEAKER_01 33:28

Yeah, we're just barely getting started, okay. Barely getting started. But tapping into the power of machine learning and AI.

SPEAKER_00 33:33

Now the other thing I want to talk about is, and I'm holding in my hands uh the CIC device that is connected directly to a smartphone. A major achievement. We're the the first to introduce a product this small that connects directly with an Android or an iPhone. And I think maybe uh, you know, people coming in from the Intel world don't appreciate fully the computing power and also the batteries have to be so small uh to in to pull this off, plus to have an antenna to transfer around from one ear to the other.

SPEAKER_01 34:06

Right. So, you know, uh I'll I'll uh let me put the engineering hat on and just uh uh explain to your your audience uh just the complexity of it. Why is it that we are the only ones with shipping this device form factor and no other companies doing it? Because it's difficult. But you also know we have a principle in uh Star key research and development that if it's difficult, we work on it right away. If it's impossible, it might take us a little longer. Right? So it is a difficult problem for all the reasons you mentioned. Uh and here you have an optimization issue. Uh you want people want their hearing aids to be as small as possible for custom devices so they don't stick out, uh they're discrete. At the same time, they want to enjoy the direct Bluetooth connectivity with iPhones and Android phones for streaming phone calls, audiobooks. And what makes that hard? Well, human body is full of water and um as a result it absorbs 2.4 gigahertz radiation very well. Yeah, Bluetooth, which is what Bluetooth is used for. There's a good reason why microwave oven that heats up our food uses 2.4 gigahertz radiation for heating up because the the water molecules in food uh absorb absorbs absorb those frequencies very well and uh resonate and uh vibrate and heat things up. Unfortunately, if you put the radio deep inside, it's gonna be absorbed by your head and no signal is gonna come out to carry the signal to your phone. So it was a very difficult problem for our radio frequency and antenna design teams, and they've done an amazing job in converting that little pull handle into a radio frequency antenna such that the device is still small and discrete, and yet we have the ability for the Bluetooth uh uh radiation to come out and connect not only with the smartphone with you to allow for direct audio streaming, but also ear-to-ear communication. Uh for you know, you want to control one hearing aid and the same instructions need to propagate to the other. So it's an example of how I have to say, in the hearing aid world, where we push the limits. Right. And you know, we are just here, I have to say, that we are continuing on a tradition of that. Mr. Austin has done that for decades. Yes.

SPEAKER_00 36:17

We have always He provided the roadmap for this by focusing on the needs of the patient, not only to customize the acoustics of hearing aids, but the form factors to make them comfortable, to make them s uh cosmetically appealing and small. And one of the other things I I think that you've really driven is uh in many cases in the past, when you chose a small form factor like this, completely in the canal device that I'm holding, or even an in-the-canal device, uh it's often the case that uh the professional feels like they have to make a concession that they can't use directional microphones because they they really won't benefit that much when you have the microphones down in the ear versus on the behind the ear ricks where they're up and they can really get different timings and to take advantage of that with directionality. But what we've done now with uh the table mic and the other accessories is enable beam-forming microphone arrays to provide very sophisticated processing, and it pairs directly with a small, cosmetically appealing device like the CIC so that the professional and the patient don't have to concede. They can have the benefits of both. They can have a small custom device that connects directly to their phone, and in noisy environments or classrooms or places of worship, they can have the benefit of that remote microphone.

SPEAKER_01 37:30

Yeah, it goes back to human perception. Uh for for uh individuals with perfectly normal hearing, we are using the the senses, the using our directional senses to cut down on uh the signals that are not interesting to us. So I'm able to g get directional signal and use my signal processing in my head, if I can say that, uh based on natural learning, not machine learning, to understand the content of the uh signal. So what this table mic does is basically provides a boost to that sense of direction when I might have lost that capability with my uh degradation in hearing. This device is amazing. It provides what, eight to ten decibels of additional uh directional benefits.

SPEAKER_00 38:20

Over what directional microphones on hearing aids alone can provide. That's eight to ten dB more benefits.

SPEAKER_01 38:26

That can it translates to it basically results in the it's a difference between not being being able to hear you in a noisy environment versus hearing you much better and understand what you're saying.

SPEAKER_00 38:37

This goes back to the superpower. It actually gives to the degree that the hearing loss allows, it gives the hearing aid user the ability to outperform their normal hearing counterpart.

SPEAKER_01 38:46

Right. So another example of how technologies like beam forming, signal processing, coupled with the in air devices with Bluetooth connectivity for audio streaming, can do miracles in terms of allowing people to perform better than that they could. In fact, you we we talked about how superhuman hearing not just have them hear better than they could, but we are on the verge of being able to enable them to hear better than even normal hearing people, which is why I use my hearing aids in challenging listening situations. Because even with my normal hearing, I wouldn't have as good a signal-to-noise ratio as I have with my devices.

SPEAKER_00 39:24

Certainly with the table mic or these uh remote microphone plus and the remote mic, giving that capability to allow the professional to choose a small form factor without conceding on performance in noisy or challenging listening environments. I want to talk about one other thing with regards to the CIC. I'll let you talk about it. But uh a really exciting feature that we've added recently for iPhone users is the ability to use the hearing aid microphone to pick up my voice when I'm talking. It's really a benefit for the person on the other end of the phone. Uh it allows true hands-free for the hearing aid users, so I'm not having to talk into my my phone mics, but for the uh person on the other end, it also provides benefits to them in better quality of my voice that they hear me.

SPEAKER_01 40:09

This was clearly uh feature that was of strong demand. And I've learned from you and others that the you know the when you had the first hearing aids in the 2012, 13, 14 that could actually stream audio from the uh iPhone and subsequently from the Android phones into the hearing aid, the value of the hearing aids went up for people because suddenly it has become their uh earbud for taking phone calls, right? Streaming uh audiobooks. Natural next step was uh just next steps in technology to have this to be bi-directional. So the hearing aids microphone can be used for picking up your phone calls so you can have the phone somewhere else, you don't have to hold it like this. But we are the first ones uh to announce the ready readiness with this technology. And you know, just just a uh a great, great testament of partnership between Apple, and I know they were Apple uh director for accessibility programs joined us at the Starkey Expo talking about how we were working on some stuff and she couldn't quite tell what are the stuff you're talking about. Well, so this is the first fruit of the collaboration that we are very happy to see in the market.

SPEAKER_00 41:19

The deep collaboration and Apple's commitment to accessibility and our knowledge in this area of those with hearing loss.

SPEAKER_01 41:25

Yep. And then we have many more things to go, and you know your audience just have to wait for us to release the next set of features. Trevor Burrus, Jr. Yes.

SPEAKER_00 41:31

It will just take a little while, a little longer, uh. Some of those impossible ones. Um the the issue with uh falls, uh I mean you talked briefly about that, but one of the things I think we're particularly proud of is that because we have the sensors in both devices, if someone falls while they're wearing them, um it does uh the the the the fall detection algorithm that sends a text and their location to trusted contacts is uh uh more accurate than the chest worn type devices. Trevor Burrus, Jr.

SPEAKER_01 42:02

So first I have to give you hard time on air, which is that I was doing a demo for fault detection and you are one of my receivers for the signal, and I get a message back from Dave. You fell again, get up. All right.

SPEAKER_00 42:13

Yeah, I'm not uh my wife and daughter insist I'm empathetically challenged. So just get up.

SPEAKER_01 42:18

But anyway, so let me get back to so the so there are a number of things that are quite amazing about our fault detection technology. Number one, the biggest value I believe is that you have hearing aids with you because you need the hearing aids to communicate with people. And now fault detection comes free. Right. You know, if you fail, the device is going to detect and alert your loved ones that you have selected on your app beforehand. You don't need to worry about those other accessories on your body that might be out of charge. In this case, you don't have to worry because you know that your your hearing aids working, it's amplifying sound, and the same device is going to keep you safe because if you fail, it would send an alert. Number two, you mentioned how we have two fall detectors. There's one on the left, one on the right. So, you know, the way that I explain it is if my left detected a fall, it would quickly check with year-to-year communication with the other one. Hey, I fell. Did you fall? So sort of like comparing notes and reducing the false positives versus one device that might be a pendant, which might flail around as I'm walking. Head is a much better place for uh detecting whether you fail or not. So it's quite amazing in my view.

SPEAKER_00 43:25

Yeah, and and of course, a fall in many cases is too late. So some uh future views is uh if if we could work to enable these devices to not only detect falls but prevent them are ways that uh uh we could expect future innovation.

SPEAKER_01 43:42

Uh your audience would be happy to know that uh we are embarking on a partnership, research partnership with Stanford University School of Medicine, where we're going to want to look at people's uh gait, their balance, because we want to reduce the incidence of falls. Today we can tell when time. Somebody fell. We really want to keep them safe, and we want to know about their deterioration in the gate before a fall happens. That would be just an amazing contribution to the world of healthcare.

SPEAKER_00 44:10

Yeah, very excited for that collaboration moving forward. So the the last thing I'll say about the existing technology, and for people interested, they can go to dot www.starkey.com and see some of the additional features that we didn't have time to get to today. But one of the things I'm particularly proud of is fall detection, edge mode that we discussed, even reminders, audible reminders for appointments or to take medication or to drink water are all cascading not only in the premier, the the premiere products, but are coming down to the other technology tiers.

SPEAKER_01 44:44

I must say, you know, the everything you said, even the just the reminder, we think of just a feature. But I was on an event presenting about it, and this they said, look, do you know how big of a problem this one is? Of the uh lack of medication compliance, where of good intentions, I forget to take my medicine, particularly people with challenges remembering things, and you have to take 12 different types of medicines in a day, and you completely forget what and when you need to take what. Now you can program your medication requirements into your into your Thrive app, that's a name of our app, and your hearing it is going to remind you at the time. Dr. Favrey, now is the time to take your take this particular medicine. That's life-saving, it's pretty amazing. This is where the third vector for hearing it's to become your personal assistant is coming to fruition.

SPEAKER_00 45:33

Yeah, and and and what do you think it's gonna take for the industry? We talk about awareness, uh, what's gonna take for the industry, the pay potential patients to recognize a hearing aid company as a healthable company?

SPEAKER_01 45:46

So I think that we are already a health company because hearing aid is a health device. It keeps you safe. I think what we what you owe to our customers and patients to quickly learn about is that we're adding a lot more value to keeping you healthy. Uh hearing better is already living better. Right. But now we're going to keep you your and your loved ones safer uh by alerting you if your loved ones are lacking social engagement. Because we know being lonely at your home, getting socially withdrawn, you are the early symptoms of dementia and Alzheimer's and cognitive declines. With machine learning capabilities, our current hearing aids, we're the only ones where we can measure your social engagement. So we can keep your loved ones safer by alerting you that you're you bought a set of hearing aids for your loved one and they are suddenly socially withdrawn. Or they used to be physically active because the built-in embedded sensors, we are the only ones measuring steps. Uh we that person was taking 10,000 steps a day, and suddenly for the last few days, they're not taking any steps. So the hearing aids now with these additional functions, what you call health-table technologies, are going to help people lead better lives more than the core function of the hearing enhancement, which already keeps them, you know, leeha leads them to lead better lives.

SPEAKER_00 47:06

Yeah, I think it's that table stakes of better hearing performance, but then the connection, as you say, to health and wellness. And there are so many uh well-published studies now that show the comorbidity between hearing loss and cardiovascular disease, cognitive decline, uh that the need for physical activity so we don't get fatter than we already are, uh, to be healthier because all of those things have high comorbidity between hearing loss. And the more that we can connect to that overall health and wellness, the better, in addition to that basic function of hearing, that the better people will perceive and again transition from half to wear into want to have.

SPEAKER_01 47:43

And I'll share with your audience uh intuition from my conversations with Mr. Austin. He's impatient. He's almost you leave the sense that we have this responsibility. We we own the, you know, a real estate in your body that's such an amazing place to deliver more value, keeping you safe. Make it an early warning system for your health issues. We just there's a sense of not only uh the desire to do things, but also a sense of responsibility of doing it. Just because we could in this form factor that you could not do in other wireable devices or other technologies. So we can't wait for this future to shape up and we're on our way.

SPEAKER_00 48:24

Excellent. Well, um it is the new year now. So in addition to the New Year's resolutions, I think the first that if someone is listening who has noticed difficulty with hearing, one New Year's resolution that you can do very easily is to get your hearing tested and see if you have a hearing loss. And there are a number of ways that you can do that through apps, through online tests. We have one at www.starkey.com to go to see a professional and and have a diagnostic test to determine whether or not hearing aids might be for you, but then now specific to your role. What are some of the new technologies that are on the horizon in 2022 and beyond?

SPEAKER_01 49:03

That's a good question. And I am just boiling inside to tell you and your audience the stories of what's coming, but I'm going to be cryptic about it. Um the devices that you have all that we already have now, with for all of the reasons that we discussed over the last half an hour, they're already just miraculously better than these devices used to be years ago. And the pace of innovation is increasing at an exponential pace, not at a linear rate. All of the things we talked about, new sensors that are being embedded, the capabilities of machine learning and artificial intelligence with enormous amounts of data that you are collecting and utilizing for developing new features or making old features better, new form factors.

SPEAKER_00 49:48

Well, the sky's the limit, and we're happy to have you at the driver's control here. And uh as they say, when you're uh given a seat on a rocket ship, you don't ask where to sit, you just sit down and uh and and get going. And uh and it's been a pleasure working with you for the past uh over three years, and I look forward to the next three years to see what we're gonna deliver.

SPEAKER_01 50:10

I have to say part of part of the the excitement at Starkey is the the talented teams we have. I'm learning every day from you on what about about the about the the human side, the the audiology, the patient's rejection or acceptance of a feature, the way the you know technologists need that because we are here to solve a specific problem or provide a specific value. So I think it's through the interactions with the users, the patients, and engineers is when miracles will happen. And you're doing such a fabulous work in bridging the tools together. Thank you.

SPEAKER_00 50:46

Oh, it's my pleasure. And um one last question, and then we'll wrap up here. Um you've mentioned uh the role of mentors in the past and how that's been important to you, but now you're in a position where you're leading large teams and mentoring them. What advice do you have for maybe engineering students, audiologists, dispensers, physicians, um based on your experiences? Do you have any any pearls of wisdom to share today?

SPEAKER_01 51:12

So I think the best I can do is just pass on. So I'll say I have through my life uh benefited tremendously from people I've looked up to, call them mentors. I still do, uh, and we'll continue to do it. Uh we never stop learning or getting inspired by people. So I would say the it's I I would say it's sticking to the basics and not overthink your entire life. But I will always challenge people to do the best of the moment and learn the most you can, make the biggest contribution you can, and then the dots connect themselves and you look back and you're amazed yourself of the path that you have taken. So just enjoy this the window of time around you in both maximizing your learning, making the biggest contribution you can, and just be nice with people around you. Uh it's about it's about having fun along this journey. And and then we can leave uh mark uh on the on the in the world if you do that in positive ways that will benefit us and benefit the society around us.

SPEAKER_00 52:12

Well, and thank you for that. And in your short time in our industry, you've already left a mark and uh we'll look forward to future innovations together. Thank you. And uh so thanks very much for being here today. Um, to our listeners, thanks for listening to this episode of Starkey Soundbites. If you enjoyed this conversation, please rate and review us on your preferred podcast platform. You can also hit subscribe so that you will be sure not to miss a single episode. And we look forward to seeing and hearing you next time. Thanks for listening.