The design of nanomaterials that directly interface with biology’s machinery of proteins represents an enormously complex engineering challenge. We typically do this by modifying single variables at a time until structure-function landscapes reveal themselves and provide design criteria. However, material cues are highly dimensional such that single variable modifications can cascade and influence each other in often unpredictable ways. This complexity manifests itself in the curse of dimensionality, making the rational design of high-performance features difficult. To overcome this challenge, we are combining active machine learning with combinatorial robotics to thoroughly survey design spaces that would otherwise be inaccessible to traditional experimentation. Employing a Design-Build-Test-Learn workflow driven by laboratory automation with active learning, we can detail structure-function landscapes with incredible precision to inform the data-directed design of biointerfacial nanomaterials.

Machine-assisted design of biointerfacial nanomaterials

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.