Polymer Screening Using Pareto-Guided Property Prediction

The system performs preliminary computational actions by training a data-driven model on historical material data and using it to predict properties of candidate materials before actual synthesis. This preliminary prediction filters out poor candidates, so that only promising materials undergo expensive and time-consuming experimental synthesis and testing, thereby resolving the contradiction between thorough search and time consumption.

The system creates a virtual copy of the material discovery process through computational modeling. Instead of exhaustively testing all possible materials in the physical world, the system uses a data-driven model to simulate and predict material properties computationally. This virtual screening copies the essential function of experimental testing at much lower time and resource cost, resolving the contradiction between thorough optimization and time loss.

The system performs preliminary computational filtering before actual material synthesis. By using a trained data-driven model to predict properties of candidate materials in advance, the system identifies and eliminates poor candidates computationally, so that physical synthesis resources are only consumed for promising candidates. This preliminary action prevents waste of substantial synthesis resources on materials unlikely to perform well.

The system replaces expensive physical experimentation with cheaper computational simulation for the screening phase. The data-driven model creates a virtual testing environment that consumes minimal physical resources compared to actual synthesis and characterization. This copying approach maintains the ability to identify optimal materials while dramatically reducing substance consumption.