The nose knows

Daniel and Chris sniff out the secret ingredients for collecting, displaying, and analyzing odor data with Terri Jordan and Yanis Caritu of Aryballe. It certainly smells like a good time, so join them for this scent-illating episode! Sponsors: Knowable https://www.knowable.fyi – Learn from the world’s best minds, anytime, anywhere, and at your own pace through audio. Get unlimited access to every Knowable audio course right now. Click here https://www.knowable.fyi to check it out and use code CHANGELOG for 20% off! Changelog++ https://changelog.com/++ – You love our content and you want to take it to the next level by showing your support. We’ll take you closer to the metal with no ads, extended episodes, outtakes, bonus content, a deep discount in our merch store (soon), and more to come. Let’s do this https://changelog.com/++! Fastly https://www.fastly.com/?utm_source=changelog&utm_medium=podcast&utm_campaign=changelog-sponsorship – Our bandwidth partner. Fastly powers fast, secure, and scalable digital experiences. Move beyond your content delivery network to their powerful edge cloud platform. Learn more at fastly.com https://www.fastly.com/?utm_source=changelog&utm_medium=podcast&utm_campaign=changelog-sponsorship. Linode https://linode.com/changelog – Get $100 in free credit to get started on Linode – Linode is our cloud of choice and the home of Changelog.com. Head to linode.com/changelog https://linode.com/changelog OR text CHANGELOG to 474747 to get instant access to that $100 in free credit. Featuring: Terri Jordan – LinkedIn https://www.linkedin.com/in/terrijordan, X https://x.com/terricojoYanis Caritu – LinkedIn https://www.linkedin.com/in/yaniscarituChris Benson – Website https://chrisbenson.com, GitHub https://github.com/chrisbenson, LinkedIn https://www.linkedin.com/in/chrisbenson, X https://x.com/chrisbensonDaniel Whitenack – Website https://www.datadan.io/, GitHub https://github.com/dwhitena, X https://x.com/dwhitenaShow Notes: Aryballe https://aryballe.comAryballe | Twitter https://twitter.com/aryballeAryballe Hardware Solutions https://aryballe.com/solutions/device-solutionsAryballe Software & Data Platform https://aryballe.com/solutions/software-data-platformHow Machine Learning in Digital Olfaction Works https://aryballe.com/how-machine-learning-in-digital-olfaction-worksReadying Odor Data For Reproduction Using Machine Learning https://aryballe.com/readying-odor-data-for-reproduction-using-machine-learningAryballe raises $7.9 million for odor-detecting AI sensors https://venturebeat.com/2020/07/10/aryballe-raises-7-9-million-for-odor-detecting-ai-sensorsUpcoming Events: Register for upcoming webinars here https://practicalai.fm/webinars!

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  So on a nominal level, a car smells like a car unless it's new or unless it has a freshener in it. We had those two outliers, one that had a cinnamon freshener and one that had a new car smell and was more recent. So now we can start to experiment with let's impart odors. And you know, some of the fun pet odors I talk about, you know, we actually have real doggy doo doo in one of the pets. So that's a day in a life of yours truly in the commercial world.

• 0:26Chris Benson?

  You're talking to a guy who has nine dogs in his house right now.

• 0:29Speaker 00

  Ah, yeah. So we can smell doggy doo doo.

• 0:32Chris Benson?

  That just made it very real for me.

• 0:37Speaker 02

  Bama for change log is provided by Fastly. Learn more at fastly.com. Our feature flags are powered by Launch Darkly. Check them out at Launch Darkly.com and we're hosted on Leno cloud servers. Get a hundred dollars in hosting credit at leno.com slash change log.

• 0:54Speaker 03

  Have you heard about Knowable? It is an awesome new platform for learning from the world's best minds anytime, anywhere, at your own pace through audio. Learn about the performance benefits of a plant-based lifestyle from NBA All-Star Chris Paul, or how to launch a startup from Reddit co-founder Alexis Ohanian. There's even a 10 lesson course from astronaut Scott Kelly. Here's a sneak peek.

• 1:19Speaker 04

  We learned a lot up there, but what can you learn from a life in space? The answers might surprise you. In this Knowable course, I want to share some of the things I've learned that you might not expect. Lessons about leadership on a dark night on an aircraft carrier in the middle of a churning sea. Lessons about the fear you feel with seven million pounds of thrust exploding underneath you. And most of all, there's an idea out there that astronauts are always perfect. Failure is not an option, right? That's why I want to take you through some of my life experiences to show you how that's just not true. I believe everyday, regular, human failure, if we handle it right, can be one of our greatest opportunities to learn,

• 2:04Speaker 04

  grow, and succeed.

• 2:07Speaker 03

  Knowable is accessible on your phone and on the web, and each audio course is broken out into individual lessons, usually around 15 minutes long. As a Change Log listener, you can get an annual membership to Knowable for 20% off. Get unlimited access to every Knowable audio course right now. Just download the Knowable app or visit knowable.fyi and use code CHANGELOG for that 20% discount. We put a link in your show notes for easy click-ins. Check out Knowable today and start learning from hundreds of top experts from around the world. Once again, that's knowable.fyi, code CHANGELOG.

• 2:51Speaker 03

  Welcome to Practical AI, a weekly podcast that makes artificial intelligence practical, productive, and accessible to everyone. This is where conversations around AI, machine learning, and data science happen. Join the community and Slack with us around various topics of the show at changelog.com/community and follow us on Twitter. We're at Practical AI FM.

• 3:17Daniel Whitenack?

  Welcome to another episode of Practical AI. This is Daniel Whitenack. I'm a data scientist with S I L International, and I'm joined, as always, by my co-host, Chris Benson, who is a Principal Emerging Technology Strategist at Lockheed Martin. How you doing, Chris?

• 3:34Chris Benson?

  I am doing very well today, Daniel. How are you?

• 3:36Daniel Whitenack?

  I'm doing well. Trying to catch up on first of year things. Lots of planning, lots of new, new things going on, new projects at work. So it's, it's an exciting time, but it's a, it's a busy time for sure.

• 3:50Chris Benson?

  It is, and it has been a crazy time out there in the world for everybody, certainly in the United States, if you're listening here. No matter what your political persuasion, it's been kind of a crazy last week. And so normally I encourage people to come to the show just for the sheer exhilaration and excitement of AI topics. But I'm hoping this, they can just come and relax because everyone I know around here is stressed out. And so we're going to have a fun time today. We're going to have a good conversation and everybody can chill out and relax and we'll talk some good AI stuff.

• 4:19Daniel Whitenack?

  Cool. Sounds good. Well, Chris, how is your sense of smell?

• 4:23Chris Benson?

  Oh, I wish it was better. So my wife, she loves scents and we have scents in the house and she'll, she'll say, what do you, and I'm like, I can't really sell it. And then she'll put it in my nose and I cough or something.

• 4:33Daniel Whitenack?

  Yeah, you're not like the sommelier, you've got the detecting the fresh grass and

• 4:40Chris Benson?

  I know.

• 4:40Daniel Whitenack?

  Rubber water hose.

• 4:43Chris Benson?

  I think this is the episode, knowing what's coming, that my wife will actually listen to, because I'm all, you know, she's like, I don't want to listen. The last thing I want to do is listen to you for a little while. So this might be the one.

• 4:54Daniel Whitenack?

  She has to do that all the time, right?

• 4:55Chris Benson?

  I know. She can't get rid of me. So yeah, this might be the episode that I pull her in.

• 4:59Speaker 00

  Super.

• 5:00Daniel Whitenack?

  Yeah, I have to confess, this one has a bunch of personal interests for myself. Our listeners might know that my wife owns a candle manufacturing business. And I know one of the things that they're always trying to do is compare different of their fragrances. They get a bunch of sample fragrances in, they analyze them. And, you know, through that I get a chance to sort of smell all sorts of things. And I don't think my sense of smell is so good. But I've had this conversation with people at work that, you know, if there was a way to like objectively or quantitatively, look at odor or smell, then that would be a huge boost and a benefit to what they're doing. And I'm really excited today that we

• 5:42Daniel Whitenack?

  have Terry Jordan and Yanis Karatou from Airball with us to talk about digital olfactory and electronic noses and all sorts of amazing things that they're doing in terms of odor analytics.

• 5:58Chris Benson?

  This is so cool, isn't it?

• 6:00Daniel Whitenack?

  It is cool.

• 6:00Chris Benson?

  We haven't done a show like this.

• 6:01Daniel Whitenack?

  Smells wonderful. So welcome, Terry and Yanis.

• 6:06Speaker 00

  Thank you. Thank you so much. Great to be here.

• 6:08Daniel Whitenack?

  Yeah. So why don't we start out before we get into this really cool topic of electronic noses and odor analytics and the amazing things you're doing at Airball. If you could just give us a little bit of a sense of your background and how you got to be interested in what you're doing now and doing what you're doing. Terry, do you want to start us out?

• 6:28Speaker 00

  Sure. So yeah, my background is actually not in anything like AI or machine learning, which is a great, great thing that Yanis is here. He's going to be my ace in the hole for this one. But I'm more of a material scientist. So my background, I have a PhD in polymer physics and had many years in big industry looking at new materials and Introducing and solving material problems. So I got involved in kind of more or less the device side. Let's say when things you can activate and make things happen i- in the solar industry. So I I started doing both startups and kind of device orientation in my commercial side. When I stopped being stopped being a scientist, or let's just say stopped being a a serious scientist

• 7:11Speaker 00

  and started being more of a serious commercial person. And so that's kind of got me interested.

• 7:15Daniel Whitenack?

  We'll still consider you a serious scientist.

• 7:17Speaker 00

  Thank you.

• 7:18Daniel Whitenack?

  You have permission from us.

• 7:20Speaker 00

  Okay, thank you. So through solar, so that that got me started through solar. And then I did UV emitting LEDs and and also a startup. And then through the company that acquired the UV LED company, I was doing portfolio and and business analysis and found Airball. So they were founded in 2014, and maybe we'll talk a little bit more about that. But they were just expanding to the US and asked me through our interactions if I wanted to run the US division. And I said, absolutely. So cool, like you said, right? So cool. Yes.

• 7:53Daniel Whitenack?

  Definitely.

• 7:54Speaker 00

  You had me at hello. So yeah, very, very excited to be here. And it's been quite an evolution for me and my career, but this is probably the most fun I've ever had in a position. So it's great.

• 8:04Daniel Whitenack?

  So cool. Yanis, what about yourself?

• 8:06Speaker 06

  Yes, on my side, I have a physics and optics background and then signal processing mostly. So I've been exploring early neural networks, structures during my PhD specifically on multisensory system. But of course, it was before year 2000. So since then, I have always been in startups for developing what we can call augmented sensor. So always sensor with something physics to translate a physical phenomenon to something in digital and then apply artificial intelligence to learn many things.

• 8:47Speaker 06

  So now I'm chief software officer at Arribal and joined the company two, two years ago. And of course, at Arribal, we have multiple skills, more chemistry, physics, optics, micro and nano electronics also we need and signal processing and of course, artificial intelligence. And that's multidisciplinary context is very exciting.

• 9:12Speaker 06

  And especially for applying this to other scenting, of course.

• 9:16Chris Benson?

  So I'm going to ask some really novice questions because I'm completely new to this. And I guess I'd like to start out by, by just kind of, if you could lay a little bit of background for those of us who are not in this industry. I know that over the years I have seen things having to do with scent. But I know that's also, you know, I'm kind of peripherally aware that it's evolving over time. Could you kind of talk about like how digital or electronic, you know, smell, that industry and those types of technologies have evolved a little bit over time? To give us a little bit of a sense on, you know, where you've arrived, you know, as an industry right now, is what the technology is looking like? Before we dive into today's technology, just give us a little history on that.

• 9:57Speaker 00

  Sure, yeah. So, you know, digital olfaction generally defined as digital capture and production of aromas. And it, you know, as you think of our odor and our sense of smell, similarly to visual, right? So as human beings, we have, you know, these beautiful slits we call eyes. And we see things in two dimensions and we form a picture in our head that makes it perceptual and turn it into three dimensions, right? And then put it back out. We do that from the time we're kids. We do the same thing with our noses, right? We have two slits and we call those nostrils and we brings in, in, and we have, there's biology in our noses and then there is electronics in our brain, right? So it really is, you know, there's a perceptual piece of this as well.

• 10:38Speaker 00

  And so unlike the visual where there's, you know, kind of clear standards and there's colors and there's RGB kind of definitions, almost none in odor. So actually, if you look at, kind of, you asked about the history. If you look at what people have tried to do or called bioelectric noses, it really begins with kind of analytical chemistry, something else called a gas chromatograph, GCMS, where you're identifying specific molecules. So it's kind of come from there. There are some general classes of molecules that have certain. Kind of let's just say earthy or musty or or skunk smell or there's there's a molecular chemistry to it, but it's not perfect, right? It's really is something that our brains have

• 11:23Speaker 00

  kind of contributed to. It's an evolutionary kind of a science. So it is something that is very well addressed with machine learning, right? We're not trying to analyze chemicals. We really are not. We we use peptides which are, and tell me if this is too specific, but peptides which are kind of like miniature proteins that are mimicking what's going on inside the nose so that we can get at the machine learning which is going on, which is kind of our brain. So that's a little bit, again, still, we're analyzing the gas, just like in the old days of GC mass spec, but we're doing it in a way that we learn, and we learn, and we learn, and we learn. And so the bioelectric nose is nothing if you don't teach it. And so we teach it using something that is mimicking the human nose, the human sense of smell. I don't know if that if that helps, but yeah.

• 12:08Chris Benson?

  It does. It does.

• 12:09Speaker 00

  So that's my answer. Yeah.

• 12:10Chris Benson?

  And just as a follow up comment to that, I love how you drew the distinction with visual because it never occurred to me. I've been programming for years. I'm so used to using RGB and stuff as a mechanism to understand that. But then, as you were talking, and I started thinking, I don't have that for all of my other senses.

• 12:27Speaker 00

  No.

• 12:28Chris Benson?

  So that's it. It's a good way of showing that gap.

• 12:29Speaker 00

  Yeah.

• 12:30Chris Benson?

  And also, you're forcing me to go back to all my intro biology chemistry classes to remember all those years ago and try to remember as you're talking peptides. And I'm like, oh, my God, I remember something about that.

• 12:40Speaker 00

  Yes, right. Right. And that is one of the things that makes us unique. I mean, there's other e-noses that are trying to mimic smell through other means, just to say, you know, solid state, you know, different kind of like electrical impulses. But we're one of the few, if not the only one, using the short chain peptides to mimic that kind of what's going on biologically as well. So would you add anything to that, Yanis?

• 13:02Speaker 06

  I can complete also, and I like this parallel with the vision also. And if the question is at the same level of what is the digital vision, the story, and try to make a parallel, it's very interesting. And what we can say is that this domain of digital olfaction is in fact very, very large. And today, all the engineering activities is really focusing on enabling the capture of others' information, thanks to specific technology of sensor which are not Always the same, but since they are comparable, because they are playing with materials, with different biosensors, like at Arribal.

• 13:43Speaker 06

  And how it works, it's, we record, of course, and store something coming from the interaction between molecules, just as the animal noise is doing. And then the question is, what we can do with that, so that would help real use cases? And here is where artificial intelligence can help. But digital olfaction may be compared just in a certain way to vision, because as it is a sense, we have now also the digital world, which is very helpful. And their respective history, vision and olfaction, along with artificial intelligence, are really not the same. So the situation in which we are right now, if you want, can be compared as if we had

• 14:26Speaker 06

  the benefits of all the artificial intelligence and all the digital world facilities we have right now. But when the photography in the past proposed to us images, when you were able maybe to differentiate people from objects, but due to the low sharpness of the images, when it was still hard, for instance, to recognize faces or to detect a red flag in a black and white picture, something like that. So I would say that we have the chance that right now we have artificial intelligence. They are almost commodities, we can say that, because of course there are progress with that being made.

• 15:03Daniel Whitenack?

  Absolutely.

• 15:04Speaker 06

  And it's a revolution for the industry. But the main difficult part of digital olfaction is maybe to grab, to capture and to capture it a more reproducible way and also displaying, because today we have screen for image, but we don't have any very sophisticated diffuser of odor, so that once you store an odor, you can restitute exactly the same, just like an image file on your hard disk and you can play it, of course, on the screen and you can recognize it even a posteriori. It is not the case with the odor. You need to say, this is smelling green grass as soon as you capture it.

• 15:41Daniel Whitenack?

  Yeah, that's really interesting. And so based on what I said at the beginning, our listeners probably know that I'm interested in this and I have some ideas in my own mind. But for those out there that are thinking... Why would we want to digitally represent odor? Like, what are the sort of main end goals or use cases that you're running across? Could you just give them a sense of the, the sort of main use cases that you're seeing right now?

• 16:10Speaker 00

  Yeah, yeah, this is what gets me really excited. So I'd love to go first on this one, too. So food and beverage industry is, I'm seeing increasingly. So when you think about, you know, raw material quality verification, think about all the places grain goes or corn, you know, different kinds of corn. So really having quality corn, we have one use case where we're looking at, where we can differentiate between good corn, musty corn, and mildewy corn, or sour corn. So looking at, so raw material supply and raw material quality verification, that goes into not just food, but also the entire fragrance industry. As you could probably imagine, we're working with a very large fragrance supplier

• 16:54Speaker 00

  that supplies to all of your consumer product goods and, you know, all of the hand sanitizers that we've been using so much of in the past year, as well as in fine fragrance. So all of that requires really fine raw material, but also product consistency on the other end. So process control. So did I switch out a batch of orange blossom for vanilla? You know, that's an extreme case. Did I bring this out of the right vat? And does this look like it's supposed to look? Those are some of the, kind of the obvious ones that we term those kind of things as cost of quality. So if you're off quality, you're costing a lot for a manufacturer of any of those, whether it's food, flavor, or fragrance. But some other ones that I think are interesting that are coming up that were surprising,

• 17:37Speaker 00

  at least to me when I started, is in the automotive industry. So in the future, and really, to a certain extent, even in the present, we're seeing more autonomous and lots of ride share, right? So ride sharing has become more or less ubiquitous in some form or another. Whether you're talking about a fleet of like, you know, Hertz, Avis type of thing, or you're talking about the Zipcar kinds of things where you're, you're actually literally sharing a same car with somebody else and it doesn't have a place to go after. So what the automotive industry is, is finding is that they need a way to take that subjective experience of getting into a car and, A, it smells bad, or it smells like new plastic, or it smells like...

• 18:19Speaker 00

  You know, somebody's had their pet in here and we need to automate that, especially when you think and you go forward into mobility and the autonomous space. There's not going to be a place to, you know, that, that subjective thing. You don't want to use your customers and clients as your sensor, as your detector. You want to sense that beforehand. So both for safety and preventative maintenance, but, but especially for the comfort and the experience of driving. All of the automotive tiers are looking at ways to standardize and create sensors that can speak back to the cloud and, and go back to wherever the kind of the controls are for that experience. So that's been really fun. We've smelled all kinds of bad odors in that case from, yeah. So we could, could share stories about that.

• 19:04Speaker 00

  That's been fun. And then consumer appliance industry as well, you know, food spoilages and food conservation has been, become a kind of a mega trend in a way. And so food storage monitoring, whether it is on a large scale supply chain of shipping food or your own refrigerator, we're looking at applications where you could imagine in the future saying to your smartphone or Alexa, whatever, to say, hey, am I out of cucumbers? I'm at the store, what do I have? And be able to set, to, to really pick up the odors of what's in your refrigerator and be able to kind of dissect that. Or you might get an alert that you're, you should be using your meat within the next couple of days.

• 19:46Speaker 00

  And so it's like detecting odors in a refrigerator for a consumer as well as in some of these larger supply chain ones. And yeah, and even automated cooking. So kind of, we did a use case where you can tell when the, when the chicken is done, when the cake is done, those kinds of applications as well. So there's a lot of spinouts from that, including, you know, lots of personal care and cosmetic examples. There's some more, a little bit more, slightly more in the future is more health examples where you can imagine wearing a, a, a watch and breathing into it and having it tell you, yeah, you should go maybe get checked for strep throat or something like that. Certain diseases have, particularly in diabetes and whatnot, have byproducts that end up in our breath or in our skin.

• 20:31Speaker 00

  So we have some, in terms of our roadmap, that's in our roadmap as well. But the first few examples I gave you are things we're working on right now with lead using early adopters in the industry.

• 20:47Speaker 03

  ChangeLog++ is the best way for you to directly support Practical AI. Join today and unlock access to a private feed that makes the ads disappear, gets you closer to the metal, and helps sustain our production of Practical AI into the future. Simply follow the ChangeLog++ link in your show notes, or point your favorite web browser to changelog.com/plusplus. Once again, that's changelog.com/plusplus.

• 21:40Chris Benson?

  So I'd like to actually talk a little bit about the actual, and I don't know, I'm going to leave it to you to define it, but the kind of the electronic nose device, if you will, that you have. And I'd really like to hear, because you both have interesting perspectives on it, if you could both kind of contribute from each of your perspectives, what that device is, both as a data producer, and as kind of from the olfactory science aspect of it. And then maybe as part of that, maybe Yanis, you could talk about kind of what data you're pulling from that, that's going in to models for development, or inference, or other software uses and stuff. But I'd really like to kind of get kind of that, both sides of what it is that you're doing.

• 22:22Speaker 00

  Great.

• 22:23Chris Benson?

  I'll leave it to you all to decide who goes first.

• 22:25Speaker 00

  Yeah. Well, we say we're hardware-enabled software data ware. So that's how we define ourselves as a company. So I'll go very briefly and say that our hardware is currently, and this year we'll be launching as a product in the market, silicon technology. So we're grafting these peptides onto a silicon chip. So it's basically kind of everything about silicon photonics that you do or don't know. It's really a silicon photonic technology stacked on a chip, where we use those peptides in a coating process. And everything else about it is just kind of wafer-level processing and pretty standard technology. The new part is kind of the peptides, which I spoke about before. So that's my short piece. The rest of it is what you do after. So I'll let Yanis go from there.

• 23:09Speaker 06

  Yes, and generally speaking, when you are talking about other information... We're entering something which is quite complex, not sure, because an odor is a complex mixture of different molecules that provoke, instead, in your nose, some activities. But the interpretation of the odor is in the olfactory cortex, so not at the same place. So we have to cope with that. And so our first focus is on delivering a sensor that will be repeatable in the term of activities, so that this information is quantified, once for all, and you inject this in the machine learning. And maybe, maybe you will mimic also the human brain,

• 23:54Speaker 06

  but not sure. So, because it depends also, interpretation of odors depends also on your history, of course, and so on. So we are applying artificial intelligence as soon as we get data from the sensors. And these sensors are just like a fingerprint sensor, because it registers patterns when you applied some odors. And this is what we try to classify, to recognize, and so on. We cite plenty of applications within different areas, but within this area, you can find recognition problems, you can find regression problems. So you can, for instance, detect which kind of odor is in this car,

• 24:38Speaker 06

  but you can also detect where is it? Where is the source of this odor? Because it's something you want to know to solve the problem. So there are plenty of opportunities to apply artificial intelligence with those outputs. So now, the sensor is providing to us a panel of figures, just numbers, which is approximately, which is less than 100 figures, which is revealing the nature of the odor. So we have plenty of pre-processing things to make it clear, to do what we call the normalization, so that you get rid of the concentration of the odor, because you have to cope with that also.

• 25:21Speaker 06

  At different concentration, you may have different signature. But we find a way also to get rid of this concentration and to access to the very nature of the odor, just like a face recognition, you need to have the face within a specific square, one by one, for instance, that's the kind of operation we are doing on the signature. And yet then we have contextual data to embed information more than the signature. So you get the, what we call the annotated database for training. And then with this annotated database, within a universe of odor, like Terry was mentioning,

• 26:02Speaker 06

  you have the possibility to create a model, which is a small brain with a small universe of odors, but maybe a big universe of odor also. And then you can recall with this model, the odor that your sensor is experiencing right now.

• 26:17Chris Benson?

  I got a quick follow-up question, because one of the things that you said along the way, I raised it and I was, I didn't want to interrupt you because what you were saying was fascinating, but I'm going to go back for a second. When you were talking about that data representation, and this may be too abstract to easily describe, but I'm kind of curious, obviously you have some sort of, the numbers are representing thing, you talked about like 100 numbers or whatever it is, is that a proprietary model? Is that something that is an industry standard that already existed? In my head, I'm thinking about like the way I tie visual to RGB when I'm programming and when I'm dealing with devices. And so I'm trying to find some sort of analogy about, aside from just going into the model,

• 26:59Chris Benson?

  how might I conceive of the representation that you have and tie that to what I'm smelling? I'm trying to make a connection between the real world and that digital, you know, representation of it that would make sense to me.

• 27:12Speaker 06

  Yes, and thank you for the question, because this is the right question, and it's not so easy.

• 27:17Chris Benson?

  Well, thank you.

• 27:18Speaker 06

  I will try to not be too much technical, but you pointed, we need to find something which is equivalent to RGB in image, or many other, like the frequency in the audio and so on. So what we are measuring are quantum of molecules with different properties, and so it's difficult to have sensors that are on a unique dimension. So collecting all this data within a specific, I would say, universe of odors to limit a bit the complexity. You can find what we call orthogonal axis in representation of data so that you have virtual sensor, which are not exactly the sensor you have, but you try to find the

• 28:00Speaker 06

  different axes which are not correlated and which put you in a world which will provide you with the less something which is not so high dimensional.

• 28:11Daniel Whitenack?

  It's kind of like principal components or something.

• 28:14Speaker 06

  Exactly.

• 28:14Speaker 00

  Principal components, exactly, yeah.

• 28:16Speaker 06

  And then this is the first step. So to define something which is like the first component, which is frozen, because you decided what are the reference odors. And then trying in the future, it's not yet there, but trying in the future to interpret the axis when you are closer to specific orders, because when we are talking about odor, we like to think about odor of flower, odor of fruit and so on. But today, it's more on the chemical point of view that we can decide when you are moving into this specific space, you are going, for instance, towards a specific chemical group or chemical family.

• 28:56Daniel Whitenack?

  Yeah, that's really helpful. I know some of our listeners may be familiar, like when when you have a machine learning problem and you have a sort of many dimensional or many different features of a problem, a lot of those can be correlated. And so that's where, you know, one of these dimensionality reducing methods, or even just orthogonality methods, like principal components come in to help you figure out what are those independent features that are going to help you describe the information of your problem.

• 29:26Speaker 00

  Yeah, the other thing to emphasize there, what Yan has talked about is reference molecules in the universe. So like you talked about RGB and we don't have that in smell. But and it's why we limit ourselves. Let's say we're going to talk about fine fragrance universe, or we're going to talk about automotive universe, which is really concerned with malodors, bad odors.

• 29:45Daniel Whitenack?

  So it's like a domain.

• 29:47Speaker 00

  And our domain, right? We define in that field or that domain, which are kind of the endpoints or the relevant ones. And we, you know, we can fine tune that as well. But being somewhat simplistic there. And then from there, we can determine how far are we away. And from there, you can start to use just regular statistics, right, to do things like confusion matrix. Are you likely to be confused that these two are the same, right? That may be a framework that your your listeners are used to. So and we can tell the difference between, I think one of the things is fascinating as a non-machine learning, non-AI person is, you know, lavender, there's, let's just say lavender, rose, is a number of things that you can take very disparate chemical structures. Again,

• 30:29Speaker 00

  being a polymer chemist, physicist myself, you can take things that are chemically looked as similar, but then they smell all like lavender. And our technology, because it's not trying to reproduce something specifically chemical structure-wise, but more that structure of interaction with your nose, that we, those tend to be grouped together in our technology. So lavender smells like lavender in our nose, in our neos, in our, in Aribal's technology, rather than in a, like the old-style GC mass spec, which is, it's only going to group together if it's chemically similar. So

• 31:03Speaker 07

  Right.

• 31:04Speaker 00

  Um, so I think it's important to kind of like, that, that concept of make sure, making sure you're defining what, what is your universe that you wanna know about, and we do that use case by use case generally. Sorry, Yanis, I maybe interrupted or co-talked over you there.

• 31:16Speaker 06

  It's okay. And, and also at Aribal, we have, of course, there are many research and development with the lab, but also we have engineering activities because it's also our mission to assess a sensor of odor. And we were talking about standardization some minutes ago, and I think one thing which is important is when talking about electronic nose, we should have a tool to assess the performance and to be able to compare to other sensors. That's a quite heavy activities, but what is specific here is that compared to other sensor I experienced in the past, is that the stimulus you have to put on your

• 31:58Speaker 06

  sensor is very combinatorial. So you have a lot of combination of different molecules to have an odor, to get an odor. It's not the same as a Having a pressure sensor, so you apply some pressure and you know where you are because you are, you have your gold. Here, it's not quite gold. The gold is indirect. So you need to specify the world of others you are playing with to be able to compare sensor to sensor and apple to apple.

• 32:27Daniel Whitenack?

  Yeah, it almost seems like there's this sort of transition where like if I think about computer vision or something like that, we're used to like on a manufacturing line, if I'm putting in, we talked about this with one of our guests, I'm putting in like a chip on a board or something. Computer vision at that point would be like, oh, I'm going to detect edges. And I know like, once I know where this edge is, then I can move my chip over and put it in in this spot. Like in that case, you're not detecting like this is the socket here. Like I'm detecting what it is. You're detecting like edges and sort of making some inference off of where it is here. You know, whereas used to you would kind of detect a certain combination of chemicals or molecules and sort of deduce like Terry,

• 33:11Daniel Whitenack?

  you were saying like, this could have to do with lavender, right? It's consistent with that, but you're not really detecting lavender smell, right?

• 33:19Speaker 00

  Correct.

• 33:20Daniel Whitenack?

  Here, it's actually more of a perception type of deal.

• 33:24Speaker 00

  Right.

• 33:25Daniel Whitenack?

  Which is really interesting. For the benefit of our listeners, I just wanted them to have in their mind this electronic nose that we're talking about. At least the one that I see in your pictures and maybe it's been updated since or this new version is coming out, but it seems almost like a sort of, if people imagine one of those older style brick cell phones about that size with a little screen on it and it's got a little, I don't know what you would call it, a nozzle orifice, something that I guess the sniffing mechanism?

• 33:55Speaker 00

  Right. That's the fluid X, right? The gases go up there, right?

• 33:59Daniel Whitenack?

  Sure, sure. Yeah. So it's got a little screen on it and you can kind of bring it around. One of the interesting things as I'm looking at that picture, I'm hearing what you're talking about and also thinking about other. Machine learning problems, like if we think about, for example, speech recognition. Speech recognition gets confused as soon as there's like multiple speakers in the environment and there's noise in the environment. Or like in the case of computer vision, I could detect, you know, maybe a person or a sign or something, but as soon as I bring in other things into the picture or apply a filter or something like that, then I'm all confused. So I'm curious, as you're sort of bringing this electronic nose around and detecting

• 34:44Daniel Whitenack?

  odors, how that sort of environment plays into the whole thing and how you can achieve that sort of robustness as you're sort of moving around. And we all know that all of what we smell all day is sort of a combination of things, not just one thing, right?

• 34:58Speaker 00

  Yeah, great. That's a great one for you, Yanis.

• 35:02Speaker 06

  So, yes, well, to me, there are two different ways to take this. Some people are using artificial intelligence just as black box. We are not, and we cannot here at Arribal, because there are so many variables also in the environment, once you are in the use case, and you don't want to learn all these variables moving. So because it's too large database, and as I said before, it's easy to have millions of cat pictures, for instance, and to annotate them afterwards. It's not easy to collect hundreds of odors or thousands of odors and to have an annotation on it. So we just need to reduce the space of the variability.

• 35:47Speaker 06

  So that's why we have a large work and large efforts on trying to keep what we measure on the same normalization space. That's the first one. And also, when it comes to source separation, as you mentioned, it with a multiple speaker for speech recognition, one way to fight against this is to find the correct equilibrium in two our chemistry between Some peptide that would be specific and some other what that would mean not specific. So this is a question also of having some improvement in the sensor. So you can look at the reality with another angle to bring more and more information. That's the simplest way to do it.

• 36:29Speaker 06

  But then if you have a more issue with the useful signal you need to practice, let's say the humidity, you are okay with having a humidity sensor also because that's commodities. So you can use a humidity sensor to solve many problems. So we have plenty of tricks on the system and still be a miniature system because all what we are talking about are chips on silicon to add this solved along our progress and development.

• 37:00Speaker 00

  And maybe for clarification for your listeners, when we talk about specific and non-specific interactions, that's particularly in regards to, let's say, the chemistry. We say that by and large the peptides have an affinity for an odor, an affinity for a gas that's going over it, but not a specific interaction. And the way to think about that is strong or weak interaction. So and not in particular to certain, any particular chemical grouping, polar, non-polar, but a specific, let's say, to fluorine or to nitrogen. So there's kind of a way we can combine some of those. And I don't know if that's clarifying possibly for some of your listeners.

• 37:38Speaker 06

  It is. And there are really two ways of coping with that. The first way is very pragmatic. Okay, what is my use case? My use case maybe can be very simple. It's quality control, I don't know. But it's a classification problem with maybe five different classes. That's very easy. So let's do it. It's okay. And the other, the other part is understanding what's happening there, what our sensor is measuring exactly, how to understand that so that we can improve also the chemistry which is inside. Everything is under our development and we improve this sensor just like the photography has done in the past.

• 38:17Speaker 00

  You know, and then taking that into the practical for a moment. So an example where this comes up for us in a use case is automotive. So in automotive, you know, we have a consortium, a group of people that we're working together towards automotive solutions. And one of our partners put together twenty something cars. And the question was, okay, so this background, what does a car smell like? Before you introduce the smell odor, what is the smell of a car? And are they all going to be the same? So we looked at like twenty something cars and did a whole range of replicates and repeat studies. And the long and the short of it is, with the exception of brand new car, which does have new car smell, and I think we're probably familiar with that odor.

• 39:00Speaker 00

  All of the cars not only had the same nominal level of intensity, they had the same, what you heard Janice call signature, this very specific pattern of this is what a car smells like. So on a kind of a nominal level, a car smells like a car unless it's new or unless it has a freshener in it. We had two, those two outliers, one that had a cinnamon freshener and one that had a new car smell and was more recent. So now we can start to experiment with, let's impart odors. And, you know, some of the fun pet odors I talk about, you know, we actually have real doggy doo doo in one of the pets. So that's a day in a life of yours truly in the commercial world.

• 39:39Chris Benson?

  You're talking to a guy who has nine dogs in his house right now.

• 39:41Speaker 00

  Ah, yes. So we can smell doggy doo doo.

• 39:45Chris Benson?

  That just made it very real for me.

• 39:49Speaker 00

  We could separate between teenage girl perspiration and aged man perspiration. So these are some of like the real things that we've done with our customers. We'll spend a day in a life at an automotive OEM and after smelling it with the digital olfaction device, my customer will say, so do you want to go in there? And I say, no, I'm good. I'm good.

• 40:12Speaker 00

  But we literally put those poor smelling things inside chambers and I kind of smell it. And those are the kinds of things that we just do and we look for differences. Just like in any statistical problem is that is they're they're different. Is it is it significant? And if we add concentrate, how many shirts do we need or pieces of shirt? Do we need to add before it is significant? And how does that compare to human panel? Not Terry panel, but another human panel that came in before or after. So those are the kinds of things that we do, just taking what Yanis was talking about and turning it into kind of practically for your listeners.

• 40:47Daniel Whitenack?

  Yeah, you need to develop some type of parallel to what they have in like self-driving with the simulated worlds and, you know, simulated real world environments so that you don't have to bring in as much doggy doo doo and smelly shirts.

• 41:05Chris Benson?

  I got all the doggy doo you need, though. If you need samples, I'll just start shipping them to you. So we're good. We're good.

• 41:11Speaker 00

  Yeah, to be clear, we there are.

• 41:12Chris Benson?

  Saturday afternoon poop scooping the yard.

• 41:15Speaker 00

  Poop scooping, that's exactly what our customer did to bring in the doggy doo. But to be clear, so we do have real those real life cases, cases. But there's also a real great body of knowledge with some of our top partners that understand the chemical makeup. So there's let's say mimetic molecules, there's molecules that mimic certain smells. And we do, we do quite a bit of testing with those things as well. So just to be clear, it's all all kind of real world testing.

• 41:41Chris Benson?

  Fair enough. I'm going to turn back to Yanis for a second because you got into something a moment ago. You were starting to talk about different uses of machine learning and different approaches. And I just wanted to, before we kind of hit a little bit of it, I'm really curious what all application? You might think of it as applications of machine learning approaches or use cases that you're using it. But anything that you haven't covered that might be interesting for listeners to hear how you're using machine learning in the process. If you've already covered everything, that's fine, too. Just thought I'd throw out the last thing to see if there's anything that's been missed.

• 42:15Speaker 06

  Yes, so you mean practically?

• 42:18Chris Benson?

  Yes. Yeah, just like different types of models, different architect, anything that you can see like, oh, for this, we do that. For this, we do that, that kind of thing.

• 42:25Speaker 06

  OK, so first and very practically, we are coding in Python for sure.

• 42:30Chris Benson?

  Yep.

• 42:30Speaker 06

  We are using some very well-known library because, as you said, it's commodities. So it's around scikit-learn and TensorFlow sometimes. As we are not really today in big data things, we are trying some deep learning. But most of the time, it's a pipeline of different processing to reduce the dimension first. So first analysis could be discriminant analysis or principal components. And then we have some very simple analysis when it is a few odors which is analyzed,

• 43:12Speaker 06

  like logistic regressions, sometimes which is working much better than some neural networks, depending on the data. So our strategy is to use bench of algorithms to classify, for instance, for recognition, and to take the best, the best one, in fact. Most of the time, to be honest, they are very comparable. It's some few percentage of accuracy if you want to compare. And sometimes for some specific universe of odors, it will not have the same algorithm that it is the best. So it really depends on the use case we have. And we have a data scientist to make this happen. And one thing also which could be a challenge is to deploy also the models.

• 43:57Speaker 06

  Because how we are talking about sensor that will embed any electronic products in consumer electronics, we need to add this model play onboard, but also with connected device, IoT, for instance. And all this logistics, software logistics is quite large to do.

• 44:17Daniel Whitenack?

  I know from our previous discussions, you have like the detection and classification piece, but Terry, you were talking about a sort of data platform or data where that kind of allows you to explore what you've detected and the sort of space of odors that you're dealing with and their signature. And I was wondering if you could talk a little bit about those. So there's maybe like a first order level of algorithms that you're doing in terms of detection and classification, but then maybe a second order level around like Clustering those together, or you mentioned detecting when something changes in a refrigerator or something like that. That seems like a more of like a time series sort of deal where you're

• 44:59Daniel Whitenack?

  that's kind of an online thing that's happening on a second order level as compared to the detection and classification. So could you describe a little bit that sort of platform and and applications?

• 45:10Speaker 00

  I think that's for you again, Yanis. I mean, yeah, yeah, those are the things that we've done and implemented in terms of proof of concept with various customers. But I think, you know, how we do that time lapse, data analysis and signal processing, that's still for you again, Yanis.

• 45:25Daniel Whitenack?

  So it sounds like that there's really sort of two levels of applying machine learning and AI in in your system. I know, Terry, you talked about the sort of data platform and sort of dashboard for people about how they're analyzing smell, the smells that they've detected. And that's like detection and classification. But there's also these other things that happen, like smells change over time. And so maybe doing time series analysis or clustering and figuring out what smells are like what other smells. Could you describe a couple of those second order applications, Yanis?

• 46:01Speaker 06

  Yes, sure. Yes, it's true that we are not in a completely constant world. The the others is evolving and so the the pattern is changing along with time. If you think about it, it's a really good information. So you have you have not a picture of the other. You have a kind of video of the other. And this video, you can apply the same way the the machine learning to have a state of when do you have to harvest fruit in a field, for instance, or you can also have this trajectory of others you can get from from cooking, for instance, to know exactly when

• 46:43Speaker 06

  you need to stop the the cooking when it's done. So there are many ways like this, what you call the second order application. Also, because we are talking a lot of time about supervised learning, but you may have application with unsupervised learning just to understand, to detect how many orders I can have in my raw materials. For instance, you apply different stocks. I don't know if it is coffee or so on, and just let the system learn the different package of orders you can get from this. And also, how many? It's not because you have 10 bags of orders that you will collect 10 bags of coffees, that you collect 10 orders.

• 47:26Speaker 06

  So we have plenty of mathematics tools to get to this kind of application. But to be honest, a large part of the application are still on detection and recognition.

• 47:37Chris Benson?

  So I guess as we start to wind up here, I've learned so much because I knew nothing about this topic before we got on. So I really appreciate you guys educating me on this. But I'm really intrigued about where this industry is going now. I have so many ideas popping into my head about ways of using this technology going forward. I'd really love it if you would both kind of wax poetic for us a little bit about your take on the future over the next few years and the exciting things that we're likely to see. And I'd like you both to address it just so we can get both of your perspectives on it, if you would. Yannis, you want to start us off and then we'll go to Thierry?

• 48:13Speaker 06

  Yes, I am fascinated by home appliance because it's close to our environment and we want to be safer when we eat something. For instance, we want to have something which is not the cooking, which would not depend on the whole system you have just by the order. Just know where you can stop your cooking and it's very facilitating for the daily life. But also, we are thinking more than that and further. I think that healthcare is of very great, of great interest. And today, the main issue is that when you exhaust an order from a body...

• 48:55Speaker 06

  Today, you need to be a dog to smell it. You need to have a very low limit of detection. So that's a challenge we have to cope with with the sensor. Because we are talking about not what we call the PPM, like the concentration, not the P P B. It's far, far lower than that sometimes. So the biological activity that exhausts some others that could witness a specific disease you have is very, very exciting. And that's something which is very promising to me for other sensing.

• 49:30Chris Benson?

  All right, thank you. Terry, you want to take it home for us?

• 49:33Speaker 00

  Yeah, I'll take it home and I'm going to give you two exams if that's okay. Because there's a very pragmatic side of me that just is super, super jazzed, especially for 2021, because we really will have an expansion in the market. So we're going, we're really introducing this solid state. We've been using, like you said, the big cellphone, the aged cellphone version. We're kind of miniaturizing everything down to silicon photonics. And so the market traction, I think, in all things, everything from coffee to fragrances to appliances and to cars, you'll start to see that building over. And I think that you will begin to see that maybe beginning in 2022, depending on what kind of field you're in. And then I find personally and professionally super energizing and exciting. So that's kind of the near term.

• 50:18Speaker 00

  But I'll end with kind of like the more or less the story of, let's say, I have three adult kids and asking them what smells evoke for them. I asked them as I was preparing for this call and they, and all of them, just coming off of Christmas, said for sure pine needles and cookies. And also, for some reason, my tomato sauce. So who knew?

• 50:41Speaker 00

  And I bring up that story, and I think that you can do this as well, and your listeners could do this as well, is smells evoke in us an emotional response, right? And this, you know, when I ask my kids about what is a smell that evokes that emotional response, they immediately know, right? So now, when you think about yourself, let's say, walking through a field or going through a hike, as I love to do, and just go. Ah, and you just wanna capture that smell. There's a very futuristic application, which is you can capture, literally capture that smell in that moment. Get your signature, bring it back and send that in to, you know, let's like Shazam for odors, right? Just send that in, say, can you please reproduce that orange blossom? Or that pine needle?

• 51:25Speaker 00

  Or that, you know, this combination of Douglas fir and roots and hummus and all of that that we find in the woods. And you can consider recreating that in a very real way. We're working with our fragrance partners with that kind of futuristic possibility as well. It's a very real possibility in future years as well. So, yeah, so those are some of the things that excite me. It's kind of the mix of the pragmatic and the like, super emotional.

• 51:54Chris Benson?

  It becomes part of the memory, not just the video on your cellphone, but the smells of that as well.

• 51:59Speaker 00

  Exactly, and there's nothing more evocative than smell to capture us in memory. There's really great science and studies about that. So really bringing, literally, or feeling literally, certainly our perception of us, bringing us right back to a place. So that kind of excites me as well. So, those are a couple of examples for you.

• 52:18Speaker 06

  And also, if I may add, maybe it's something personal, but I am always thinking about a hope. Something which is a hope that the human being is progressing conversely to animals. It's well known that we forget this sense of odor. My hope is that with such a systems and very spread system, when you have plenty of consumer products that could embed this sense again. Maybe it will take sometimes, maybe, but we will have again the sense of smell, which is important. To feel, to have sensation, emotions, and so on.

• 52:56Speaker 06

  And I think it's also not a mission, but it may be a benefit in the future that we are focusing on this specific digital sense.

• 53:08Daniel Whitenack?

  Yeah.

• 53:09Speaker 00

  As a way to connect us.

• 53:10Daniel Whitenack?

  Yeah, that's wonderful.

• 53:12Chris Benson?

  It's a beautiful way to finish there. I love that. That's a great.

• 53:15Daniel Whitenack?

  So wonderful. Yeah, this has been a fascinating conversation and I appreciate you both taking time out of your busy days to join us. It was, it was really wonderful and I'm really excited to follow all the things you're doing. We'll, we'll give some links in our show notes about all of the great things you're doing and where people can find out more and encourage people to check it out because it's, it's really fascinating stuff. Thank you both.

• 53:40Speaker 00

  Super. Dan and Will connect offline, right?

• 53:42Daniel Whitenack?

  Yep, exactly.

• 53:43Speaker 00

  Okay, your other business. Okay, great. Thanks so much. It's been really fun, guys. Thank you.

• 53:49Daniel Whitenack?

  Thank you.

• 53:50Chris Benson?

  Thank you very much.

• 53:54Speaker 03

  Thank you for listening to Practical AI. If this is your first time, make sure you subscribe so you don't miss a thing. Head to practicalai.fm to subscribe or find us in Apple Podcasts, Spotify, or wherever you listen to podcasts. And if you get value from the show, please do share it with a friend or a colleague. We appreciate you spreading the word. Practical AI is hosted by Daniel Whiteneck and Chris Benson. It's produced by Jared Santo and our music is provided by Breakmaster Cylinder. We are brought to you by some awesome sponsors. Shout out to Fastly, Linode, and Launch Darkly. That is our show. We hope you enjoyed it and we'll talk to you again next week.