Artificial intelligence (AI) isn’t magic, it’s just really complicated math, said Greg Mulholland, CEO and founder of Citrine Informatics (Redwood City, CA), at a press roundtable hosted by Lanxess (Cologne, Germany) at K 2019. But Mulholland’s hosts seemed quite bedazzled by his AI-enabled platform, nonetheless. Lanxess is the first company to adopt Citrine's technology at scale, and Dr. Markus Eckert, Senior Vice President, Head of Business Unit Urethane Systems at Lanxess was eager to explain what it means for customers.
Citrine is a Silicon Valley startup that couldn’t be more niche: It has developed a platform that leverages data and AI specifically to accelerate the development of materials and chemicals. Citrine has been recognized for technology innovation by the World Economic Forum as a Tech Pioneer, and collaborates with world-class academic institutions such as Carnegie Mellon University in Pittsburgh and the University of California, Berkeley. Lanxess is leveraging Citrine’s technology to expand its range of prepolymers and speed up the development of customer-specific polyurethane systems for existing as well as entirely new applications.
Initially, the Urethane Systems business unit of Lanxess is using the platform to enlarge its database of prepolymer-based formulations. Lanxess data specialists and process experts leveraged the Citrine platform for AI to add further data points to the company's formulation database. The process involved linking existing empirical measurement data with in-house expert knowledge and a chemistry-aware algorithm to calculate additional measurement values, explained Lanxess in a press release. This meant that only a few real-world measurements were required to verify figures determined via AI. “That part of the project started just a few months ago,” said Eckert. “But the progress has been surprisingly fast, and it really works!”
The next step involves verifying the predictive performance of the algorithm to optimize formulations to meet customer-specific requirements for product characteristics. “If the next tests are successful, we will be able to fulfill customer requests even more quickly and effectively," said Eckert. He believes that AI will help the company to evaluate urethane systems that are not yet part of the company’s portfolio and determine whether or not they can be manufactured, and how, at unprecedented speed.
Prior to this announcement at the K, Lanxess collaborated with Citrine Informatics in a pilot project using AI to optimize the sizing of glass fibers in high-performance plastics.
|Citrine CEO Greg Mulholland (left) and Markus Eckert (center) and Axel Tuchlenski of Lanxess discussed the use of AI to accelerate materials formulation at K 2019.|
Mulholland stressed that Lanxess is developing AI as an “internal capability” with the "support" of Citrine. “This is not a consultancy-type relationship,” he said, and AI should not be seen as a replacement of the scientific trial-and-error process, he added. Rather, it is a time compression tool. Ultimately, AI will be part of the scientific toolbox at Lanxess and the scientists may not even realize they are using AI in their work.
“We have more than 100,000 formulations and a deep base of knowledge,” said Eckert. “AI helps us to put that treasure to effective use.”
Top image: Sergey Tarasov/Adobe Stock