(62b) Predicting Polymer–Surface Adhesion Strength Via Machine Learning and Transfer Learning

Polymer-surface interactions are crucial to many biological processes and industrial applications. Here we propose a machine-learning method to connect a model polymer's sequence with its adhesion to decorated surfaces. We simulate the adhesive free energies of 20,000 unique coarse-grained 1D sequential polymers interacting with functionalized surfaces and build support vector regression models that demonstrate inexpensive and reliable prediction of the adhesive free energy as a function of the sequence. Our work highlights the promising integration of coarse-grained simulation with data-driven machine learning methods for the design of functional polymers and represents an important step toward linking polymer compositions with polymer-surface interactions.

Building upon this foundation, we address a notable challenge: the necessity of extensive datasets for each specific surface to adequately train the surrogate models. Ideally, one should be able to infer information about similar surfaces without having to regenerate a full complement of adhesion data for each new case, making it possible to explore scenarios of insufficient data. To bridge this gap, we demonstrate a transfer learning technique using a deep neural network to improve the accuracy of ML models trained on small datasets by pre-training on a larger database from a related system and fine-tuning the weights of all layers with a small amount of additional data. The shared knowledge from the pre-trained model facilitates the prediction accuracy significantly on small datasets. We also explore the limits of database size on accuracy and the optimal tuning of network architecture and parameters for our learning tasks. While applied to a relatively simple coarse-grained polymer model, the general lessons of this study apply to detailed modeling studies and the broader problems of inverse materials design.