TransPolymer: a Transformer-based language model for polymer property predictions

Abstract Accurate and efficient prediction of polymer properties is of great significance in polymer design. Conventionally, expensive and time-consuming experiments or simulations are required to evaluate polymer functions. Recently, Transformer models, equipped with self-attention mechanisms, have exhibited superior performance in natural language processing. However, such methods have not been investigated in polymer sciences. Herein, we report TransPolymer, a Transformer-based language model for polymer property prediction. Our proposed polymer tokenizer with chemical awareness enables learning representations from polymer sequences. Rigorous experiments on ten polymer property prediction benchmarks…

• UD
  U.S. Department of Energy
• CM
  Carnegie Mellon University
• AR
  Advanced Research Projects Agency - Energy

  Award: DE-AR0001221
• AR
  Advanced Research Projects Agency