Using AI to speed up the scientific process is a concept referred to as "self-driving labs,'' applying the ideas behind self-driving cars to laboratories
A startup company formed based on technology developed at Rutgers seeks to use artificial intelligence and automation to help researchers advance their work faster and make breakthrough discoveries.
"Currently in labs, a scientist follows the conventional scientific method where they come up with a question, develop a hypothesis, design an experiment to test that hypothesis, implement the experiment, analyze the results, and then update their hypotheses," said Rutgers associate professor of Biomedical Engineering Adam Gormley, who developed the technology with his former student, Matthew Tamasi.
"The question we had was, is that the most effective utilization of a scientist's time? We thought that maybe outsourcing some of the work such as designing experiments, performing the experiments, and then doing routine analyses may ideally be suited for an automated system, an AI model. Then, the AI model could predict the next best experiment to answer fundamental questions or design material problems."
Using AI to speed up the scientific process is a concept referred to as "self-driving labs.'' The idea stemmed from the notion of self-driving cars, except using laboratories, Gormley said.
The technology, being advanced by Plexymer, Inc. was created by Gormley and Tamasi in the Gormley lab. Tamasi now serves as chief technologyoOfficer of the company.
"My lab is incredibly excited by the potential of self-driving labs, and we are neatly in the domain of applying these ideas into the fields of biomaterials, drug delivery, and polymer chemistry," Gormley said.
Polymers are large molecules that are used as the basis for many organic and synthetic materials, including therapeutics. According to Gormley, pharmaceutical and consumer products use a standard set of polymers for their applications, no matter how complex the application is, and he believes custom designed polymers would work better. The challenge, however, is how to discover those custom polymers. That is where Plexymer, with its AI and machine learning technology, comes in.
