Machine learning as a tool to design glasses with controlled dissolution for healthcare applications

Area of Science:

• Materials Science
• Biomedical Engineering
• Computational Science

Background:

• Glass science is crucial for healthcare innovations, but designing glasses with specific degradation rates is challenging.
• Conventional methods struggle to create bioactive glasses with tailored ion release for patient-specific needs.

Purpose of the Study:

• To develop an artificial intelligence-based method for designing novel glasses with controlled ion release.
• To present an ensemble machine learning (ML) model for predicting the dissolution behavior of oxide glasses in biomedical applications.

Main Methods:

• Utilized a comprehensive database of over 1300 glass dissolution experiments for training and testing.
• Developed and applied an ensemble machine learning model to predict time- and composition-dependent glass dissolution.

Main Results:

• The ML model reliably predicts the chemical degradation of various oxide glasses.
• Accurate predictions were achieved across a wide pH range (3-10), simulating physiological conditions.

Conclusions:

• The developed ML model enables the design of biomedical glasses with predictable dissolution rates.
• This approach facilitates the creation of glasses with controlled ion release for targeted therapeutic effects in diverse biological environments.