Artificial intelligence has been making a difference in the discovery lab for a decade. Recently, the BCG advisory group suggested that AI plays a central role in identifying more than 150 current drug candidates, 15 of which are in trials.
And, according to a separate Deloitte analysis, AI approaches have also been used in areas such as clinical trials, supply chain management and commercial launch.
But in the biofabrication plant, it’s a different story. Cases of AI being used in drug production are rare. And that’s a missed opportunity, according to Kiefer Eaton of industrial artificial intelligence developer Basetwo AI, particularly because drugmakers already have many of the things they need to benefit from AI.
“AI has been widely adopted for drug development and has recently seen strong growth for commercial pharmaceutical activities. However, biomanufacturing remains underserved despite the plethora of data to mine,” he says. “AI offers a unique opportunity in manufacturing through the creation of “hybrid process models” using physics-based AI. These hybrid models combine the power of engineering insights with machine learning to create digital twins, virtual representations of physical processes or unit operations.
Eaton cites the ability to optimize processes in a virtual environment without consuming resources as a potential application.
“Applying AI in process development requires two things: process knowledge and data. Leveraging the process through engineering equations allows us to contextualize the training of the model that learns to from the data,” he continues. “That means we can use less data than we would otherwise need to train a machine learning model, because process knowledge fills in a lot of the gaps. It even allows us to use data from related processes at different sites or at different scales. »
For example, data from a fed-batch process can be used to train an infusion bioreactor model by modifying the equations, notes Eaton, who adds, “This greatly improves process cross-learning for manufacturers. »
The pharmaceutical industry therefore has the data it needs to use AI in manufacturing. But building an industrial AI architecture isn’t just about data, Eaton points out, suggesting that drugmakers will need to invest in internal training and external expertise.
“A great challenge to bring the right stakeholders and subject matter experts together on a common platform to leverage the strengths of both groups. For example, a data scientist might ask Python to create a predictive process model using AI, while an engineer would stick with MatLab or other traditional modeling software,” explains he. GEN.
“Beyond that, manufacturers sometimes lack the ability to deploy AI at scale, which completely impedes adoption, especially if there is a lack of data science expertise to manage. these AI models throughout their life cycle.”
Having technology that minimizes coding requirements is also a good idea, according to Eaton, citing Basetwo’s own AI system, which has already been deployed by a Johnson & Johnson Innovation team, as an example.
“We built Basetwo from the ground up with engineers in mind. An engineer can create a data pipeline using our drag-and-drop interface, visualize the data in a spreadsheet view, and create a hybrid model using our drag-and-drop interface. using our equation editor with Excel-like syntax, all without writing a single line of code,” he says. “It makes unlocking the power of hybrid modeling and AI accessible to engineers and improves collaboration between data scientists and engineers.”