When medical companies make pills and tablets, they need to extract the active ingredient and dry it. Currently, this process requires a human operator to monitor and agitate the material, relying on their observations to determine when it’s ready for compression. However, researchers at MIT and Takeda have developed a new technique that combines physics and machine learning to improve this process.
The technique, known as physics-enhanced autocorrelation-based estimator (PEACE), uses lasers to illuminate particles in a mixture during filtration and drying. By analyzing the interaction between the laser and particles, as well as using machine learning to characterize particle sizes, the process becomes more efficient and accurate. Unlike traditional methods, this approach doesn’t require stopping and starting the process, making it more secure and efficient.
One of the advantages of the PEACE mechanism is that it doesn’t require a large amount of training data for the machine learning algorithm. The physics-based approach allows for speedy training of the neural network, reducing the need for extensive datasets.
Currently, there are no inline processes for measuring particles within a powder during mixing in the pharmaceutical industry. This new technique could revolutionize pharmaceutical manufacturing by making it quicker, more efficient, and safer. By reducing the number of experiments needed, companies can produce drugs more efficiently and cost-effectively.
This collaboration between MIT and Takeda has already resulted in 19 projects focused on applying AI and machine learning to healthcare and the medical industry. The close proximity of Takeda’s lab to MIT’s campus has allowed for real-time collaboration and feedback, shortening the timeline from academic research to industrial implementation. The team has already filed two patents and plans to file for a third, showcasing the real-world implications of their work.
Overall, the PEACE mechanism has the potential to revolutionize pharmaceutical manufacturing, improving efficiency, sustainability, and cost-effectiveness. Its applications could extend beyond just pills and tablets, making it a valuable tool for various industrial pharmaceutical operations in the future.