Webinar Recap: How Applied Molecules Accelerated Innovation with Albert
This post is a recap of our recent webinar with Applied Molecules. You can view the full webinar recording here.
Albert’s mission: To help scientists invent faster
Albert Invent was founded on a vision of the “chemist of the future” – a future in which scientists could invent faster and accelerate their impact on the world. We recently had the pleasure of hosting a webinar with one such scientist, Cameron Darkes-Burkey, Senior Scientist at Applied Molecules and someone who, in his words, spends most of his workday in Albert. In this webinar, Cameron shared the story of how Applied Molecules brought Albert into their workflow, accelerating the development of dampening materials from months to days and exploring entirely new combinations of materials for 3D printing resins.
The webinar began with an overview of Albert by CEO and co-founder Nick Talken, followed by a deeper dive into Albert Breakthrough, our suite of artificial intelligence and machine learning (AI/ML) tools, trained on more than 15 million molecules. Sebastian Bernasek, Senior Machine Learning Engineer at Albert, walked through the basics of Breakthrough at the start of the webinar: with Breakthrough, scientists can build custom ML models based on historical experimental data, generate new candidates with desired properties, and use their domain knowledge to choose which candidates to test. The new data that is collected can be fed back into the model in an iterative active learning process that simultaneously leverages domain knowledge and empirical data.
Rapid development of dampening materials: From 3 months to 2 days
Applied Molecules, a thermoset materials science company for the industrial sector, adopted Albert in 2023. “For a small, fast-moving team, Albert provides an all-in-one solution that allows us to maximize the impact of every chemist,” Cameron shared.
Cameron illustrated the impact of Albert Breakthrough for two recent projects. In the first example, Applied Molecules wanted to optimize two parameters – storage modulus and dampening factor – for a dampening material requested by a customer. Cameron and his colleagues used Albert Breakthrough to explore four main raw ingredients. The team started by entering 22 existing formulations and target ranges into Breakthrough, which suggested thousands of candidates to explore. The team chose 10 candidates to test per iteration, feeding their results back into Breakthrough and retraining the model each time. After just three iterations, Applied Molecules was able to create materials that matched the specifications requested by their customer. Typically, a project like this would have taken Applied Molecules three months to complete. With Albert Breakthrough, they were able to achieve this in just two days.
Typically, a project like this would have taken Applied Molecules three months to complete. With Albert Breakthrough, they were able to achieve this in just two days.
Cameron and Nick discussed how this workflow was different from the conventional approach. While it would have been possible to optimize this dampening material through the traditional scientific method, it would have taken much longer. He also pointed out that the increase in speed not only helps get products to customers faster but also helps Applied Molecules stay agile in the development process, enabling them to pivot to new customer specifications on the fly while still using previously collected data to inform new formulations.
Multi-dimensional optimization of novel 3D printing resins
The second use case that Cameron shared was more complex. Here, Applied Molecules wanted to simultaneously optimize five parameters (heat deflection temperature, impact strength, ultimate tensile strength, elongation, and tensile modulus) for a 3D printing resin. “This is a powerful use case,” Nick commented. “It speaks to what the members of the audience are doing every day – high dimensional optimization where you have trade-offs between multiple properties. It’s very rare that you have all five properties that move in the same direction at once.”
To tackle this challenge, Cameron and his team explored 13 new materials by starting with 60 formulations. By testing three suggested formulations per iteration, the team arrived at a successful result after 12 iterations.
What is unique about this example was that Albert didn’t just accelerate innovation – it expanded the design space in ways that even an experienced scientist may not have thought to try, a benefit that is especially useful in optimization problems that involve multiple parameters. “In this situation, we’re able to not only see Albert learn, but we’re able to learn ourselves as we’re surprised by what it’s suggesting,” Cameron reflected. “And when we run that formulation, print it, and see the results – we would never have guessed that those two materials in combination provided the results that it did.”
Sebastian also pointed out that the scientist still plays a critical role throughout the process, such as in imposing boundaries, making technical decisions throughout the iteration process, and testing. AI is not going to replace the scientist, but Breakthrough is a user-friendly tool that allows scientists like Cameron to leverage AI to innovate faster and produce higher quality products.
Together, let’s invent the future of chemistry
Cameron concluded the webinar by summarizing how Albert’s integration into the workflow at Applied Molecules has streamlined research and collaboration. We especially appreciated his comment about how there is some level of change management that is necessary, but that it is worth it for how invaluable Albert has become in their R&D processes.
We were pleased to see so many questions from the audience, including several about IP and data security, the different applications that Breakthrough can be used for, and topics that weren’t covered in detail in this webinar such as our automated SDS labels.
Speaking with Cameron about his successes at Applied Molecules left our team feeling energized about the future of materials science innovation. AI is here, it's fundamentally changing how science is done, and we want to help you start that journey. If you want to invent faster like Applied Molecules and become a leader in leveraging AI, request a demo today! We'd love to talk.
This post is a recap of our recent webinar with Applied Molecules. You can view the full webinar recording here.
Albert’s mission: To help scientists invent faster
Albert Invent was founded on a vision of the “chemist of the future” – a future in which scientists could invent faster and accelerate their impact on the world. We recently had the pleasure of hosting a webinar with one such scientist, Cameron Darkes-Burkey, Senior Scientist at Applied Molecules and someone who, in his words, spends most of his workday in Albert. In this webinar, Cameron shared the story of how Applied Molecules brought Albert into their workflow, accelerating the development of dampening materials from months to days and exploring entirely new combinations of materials for 3D printing resins.
The webinar began with an overview of Albert by CEO and co-founder Nick Talken, followed by a deeper dive into Albert Breakthrough, our suite of artificial intelligence and machine learning (AI/ML) tools, trained on more than 15 million molecules. Sebastian Bernasek, Senior Machine Learning Engineer at Albert, walked through the basics of Breakthrough at the start of the webinar: with Breakthrough, scientists can build custom ML models based on historical experimental data, generate new candidates with desired properties, and use their domain knowledge to choose which candidates to test. The new data that is collected can be fed back into the model in an iterative active learning process that simultaneously leverages domain knowledge and empirical data.
Rapid development of dampening materials: From 3 months to 2 days
Applied Molecules, a thermoset materials science company for the industrial sector, adopted Albert in 2023. “For a small, fast-moving team, Albert provides an all-in-one solution that allows us to maximize the impact of every chemist,” Cameron shared.
Cameron illustrated the impact of Albert Breakthrough for two recent projects. In the first example, Applied Molecules wanted to optimize two parameters – storage modulus and dampening factor – for a dampening material requested by a customer. Cameron and his colleagues used Albert Breakthrough to explore four main raw ingredients. The team started by entering 22 existing formulations and target ranges into Breakthrough, which suggested thousands of candidates to explore. The team chose 10 candidates to test per iteration, feeding their results back into Breakthrough and retraining the model each time. After just three iterations, Applied Molecules was able to create materials that matched the specifications requested by their customer. Typically, a project like this would have taken Applied Molecules three months to complete. With Albert Breakthrough, they were able to achieve this in just two days.
Typically, a project like this would have taken Applied Molecules three months to complete. With Albert Breakthrough, they were able to achieve this in just two days.
Cameron and Nick discussed how this workflow was different from the conventional approach. While it would have been possible to optimize this dampening material through the traditional scientific method, it would have taken much longer. He also pointed out that the increase in speed not only helps get products to customers faster but also helps Applied Molecules stay agile in the development process, enabling them to pivot to new customer specifications on the fly while still using previously collected data to inform new formulations.
Multi-dimensional optimization of novel 3D printing resins
The second use case that Cameron shared was more complex. Here, Applied Molecules wanted to simultaneously optimize five parameters (heat deflection temperature, impact strength, ultimate tensile strength, elongation, and tensile modulus) for a 3D printing resin. “This is a powerful use case,” Nick commented. “It speaks to what the members of the audience are doing every day – high dimensional optimization where you have trade-offs between multiple properties. It’s very rare that you have all five properties that move in the same direction at once.”
To tackle this challenge, Cameron and his team explored 13 new materials by starting with 60 formulations. By testing three suggested formulations per iteration, the team arrived at a successful result after 12 iterations.
What is unique about this example was that Albert didn’t just accelerate innovation – it expanded the design space in ways that even an experienced scientist may not have thought to try, a benefit that is especially useful in optimization problems that involve multiple parameters. “In this situation, we’re able to not only see Albert learn, but we’re able to learn ourselves as we’re surprised by what it’s suggesting,” Cameron reflected. “And when we run that formulation, print it, and see the results – we would never have guessed that those two materials in combination provided the results that it did.”
Sebastian also pointed out that the scientist still plays a critical role throughout the process, such as in imposing boundaries, making technical decisions throughout the iteration process, and testing. AI is not going to replace the scientist, but Breakthrough is a user-friendly tool that allows scientists like Cameron to leverage AI to innovate faster and produce higher quality products.
Together, let’s invent the future of chemistry
Cameron concluded the webinar by summarizing how Albert’s integration into the workflow at Applied Molecules has streamlined research and collaboration. We especially appreciated his comment about how there is some level of change management that is necessary, but that it is worth it for how invaluable Albert has become in their R&D processes.
We were pleased to see so many questions from the audience, including several about IP and data security, the different applications that Breakthrough can be used for, and topics that weren’t covered in detail in this webinar such as our automated SDS labels.
Speaking with Cameron about his successes at Applied Molecules left our team feeling energized about the future of materials science innovation. AI is here, it's fundamentally changing how science is done, and we want to help you start that journey. If you want to invent faster like Applied Molecules and become a leader in leveraging AI, request a demo today! We'd love to talk.