[ad_1]
Karim Aly is CEO of Noze, a Canadian AI startup that has developed world-class technology to digitize the sense of smell. He is focused on executing the company’s vision of transforming healthcare by empowering machines with the ability to smell.
Prior to joining Noze, Karim founded one of Canada’s first startup studios in partnership with one of Canada’s largest universities. Early in his career, he was an active entrepreneur in emerging markets, having founded multiple technology companies that scaled to more than 20 countries in the Middle East and Southeast Asia.
The idea spark for digital smell was first conceived in 2014, but could you share some insights from those early days?
of course. It was due to the natural curiosity of its founder and his CTO, Ashok Prabhu Masilamani. He was driven to understand what made it possible to digitize sound (microphones), vision (cameras) and touch (tactile sensations), but it didn’t. odor. As he peeled back the layers, he began to understand the main failures that have held us back in our quest to digitize smells. As a career scientist, these learnings formed the basis of his Ashok vision for new startups. Developing a platform that can truly bring the perception of smell into the digital world, he created Noze.
The company spent the next six years innovating and perfecting the world’s most advanced digital odor recognition framework that solves real-world odor detection and tracking. While this technology obviously has potential applications in a variety of fields, from air pollution to law enforcement, we chose to focus on applying our digital olfactory platform only to healthcare. In fact, they just announced a $1 million grant from the Bill & Melinda Gates Foundation to develop AI that can detect infectious diseases such as malaria and tuberculosis via odor biomarkers (volatile organic compounds) in breath. We are developing a medical sobriety detector that uses this. This will be a game changer for millions of people.
In 2015, NASA’s Jet Propulsion Laboratory (JPL) had technology that matched the team’s vision. What is this technology and how did your team secure this patent?
In 2014, NASA’s Jet Propulsion Laboratory developed a revolutionary “digital nose” technology to detect multiple vapors/gases on orbital vehicles in space. NASA has focused on testing this capability on the International Space Station (ISS). The ISS is a much more difficult environment to “smell” steam than it is here on the ground. We saw great potential in their early learning, so we decided to accelerate our journey by securing an exclusive license to his six patents held by JPL in the Digital Nose field. bottom. Since then, we have fundamentally evolved JPL’s digital nose technology by adding layers of proprietary aroma data engineering and sensory AI algorithms to launch the world’s most powerful digital odor recognition platform. , has improved.
What are the different machine learning technologies used to generate unique digital centprints?
Machine learning is not the only way to generate interpretable digital centprints. At Noze, we recognized early on that digital olfactory should be viewed as a framework similar to the mammalian olfactory system. In mammals, the front end of the olfactory system is diverse olfactory receptors. To emulate these olfactory receptors, we constructed sensor chips with diverse chemoreceptors. When odors are introduced into mammalian olfactory receptors, they generate unique neural codes. Similarly, when an odorant passes through a chemoreceptor array, it produces a unique ‘digital centprint’.
The sensory front end of the Digital Olfactory Framework is just the tip of the iceberg. It is powered by a cloud-based curated digital odor library and a chemically perceptive AI engine. Magic happens when all the pieces work in harmony.
Can you explain the algorithm used to interpret the scent print?
To interpret an odor, we need to create a dataset of digital centprints of that odor. We found that the odor datasets constructed from the Noze sensor chip contain rich chemical semantic information represented in the form of manifolds. In the world of computer vision, a common approach is to use diverse learning techniques. However, unlike computer vision, which is a data-rich area, the world of digital olfactory is data-poor. As such, our AI toolbox applies a variety of novel approaches such as meta-learning, few-shot learning, and manifold learning to a dedicated odor dataset.
Digital centprints of real-world odors typically include all relevant background noise that prevents correct interpretation. This is why our proprietary dataset is carefully curated and constructed using a combination of data points representing background odors (noise) and data points representing the odors themselves. This trains the AI algorithm to recognize and reject background noise, allowing it to correctly interpret incoming scent prints.
Tell us about Noze’s cloud-based platform, the process of adding new scents, and the size of your library of scent prints.
Our cloud-based IoT platform hosts a digital odor library and a perception AI engine. Our library consists of two types of datasets. One is actively designed to create prints of selected odors and background odors, and the other is from continuous sampling done by devices in the field, including our sensor chip. It was created passively. These passively sampled scent prints are curated and stored in an odor library so that the platform can match and match them against potential future learned odors. Given that our platform is connected to all devices in the field, we have also developed a strong network effect where there is a continuous collective learning process across devices. In other words, one device can learn to interpret new odors from learning acquired by an entirely different device.
We made a fundamental decision to focus on building high-quality scentprints that enable meaningful use cases. It is based on the economic and societal value that can be unlocked from the scent library in the world. Includes over 100 scented scent prints.
What are the different use cases for digital centprints in manufacturing?
It’s easy to start imagining how nearly every industry could benefit greatly from the digitization of the sense of smell. In manufacturing, there are some clearly valuable use cases, especially those related to improving safety and ensuring regulatory compliance. Imagine being able to detect burning wires in a machine just by the odor emitted, thus giving you the opportunity to stop operation before a fire starts. Or continuously track the collection of by-product vapors to evict and ventilate the area the moment concentration crosses his HS&E threshold.
The unique ability to distinguish odor signals from background noise allows you to determine that the actual source of the odor is coming from an actual burning wire, not something like cigarette smoke or hot coffee. Avoiding false positives due to other “background” odors is critical and one of the biggest challenges for the successful commercialization of a digital olfactory platform.
How is this technology currently being used with food?
Our technology is not currently used in the food industry, but it has many potential applications throughout the food supply chain. Take food freshness as an example. What if you could detect the food items in your refrigerator and predict how long each food item will take to spoil? This same solution can be applied to grocery stores and restaurants. Together with households, he accounts for over 80% of the food wasted each year. This is his $400 billion problem in the US alone.
From a completely different angle, digital smell can also help automate the cooking process by tracking the aromas of dishes and recipes from start to finish. This allows you to tell the chef (or automate the utensils) what to do at each step of cooking. the way. In fact, he created a demo to train an AI on the process of perfectly cooking a chicken breast on an indoor grill. I was able to let the user know when the grill was hot enough to add the chicken, when to flip it, and when to remove it from the grill to complete a perfectly cooked chicken breast.
One interesting use case is virus detection. Please let me know how this works.
The human body releases specific odor biomarkers, or volatile organic compounds (VOCs), as a physiological response to infection. However, this phenomenon is not limited to viral infections. These VOCs can be emitted from our breath and skin and can indicate the presence of various clinical conditions and diseases. When we think of “healthy sobriety detectors,” which could detect malaria, tuberculosis, diabetes, and other conditions in their early stages in a single breath, it’s easy to understand the impact of our technology on drinking ability. can do. Timely action and improved patient outcomes. It is this vision that we are currently working with several partners, including the Bill & Melinda Gates Foundation and the Montreal Heart Institute. As a company, this is where we found our sense of purpose, and we couldn’t be more excited about both the work we do and the meaningful impact it makes.
What is your vision for the future of digital olfactory recognition?
Noze’s digital olfactory platform is a powerful digital olfactory tool. Over the past eight years, we have perfected this technology to work outside of a controlled lab environment. We have built several odor detection or tracking solutions for everyday scenarios. Our solutions work reliably despite the challenges associated with each. Our goal today is to apply this technology to take human health to a whole new level. We’ve only scratched the surface in terms of what we can interpret from the volatiles we’re continuously emitting from our breath and skin. We believe that correlating presence with various health conditions could dramatically change the medical landscape. . Volatiles of interest are usually present with a confounding background such as the presence of exogenous VOCs, high temperatures, and condensed humidity. Each of these characteristics can affect detection accuracy, making it particularly difficult to build a reliable and scalable solution.
Our vision for Digital Smell has therefore always been clear: to provide a robust and scalable solution that works reliably in the real world as well as in the lab. Only then can we truly enable ubiquitous access to screening and diagnostics that can help save lives and improve health. And today we are bringing it to the world.
[ad_2]
Source link