Polymer-Agent: Large Language Model Agent for Polymer Design

On-demand polymer discovery is essential across various industries, from biomedical applications to reinforcement materials. Experiments with polymers involve a long trial-and-error process that consumes extensive resources. For these processes, machine learning has accelerated scientific discovery on the property-prediction and latent-space search fronts. However, laboratory researchers cannot readily access codes, and these models to extract individual structures and properties due to infrastructure limitations. We present a closed-loop polymer structure-property predictor integrated in a terminal for early-stage polymer discovery. The framework is powered by LLM reasoning to provide users with property prediction, property-guided polymer structure generation, and structure modification capabilities. The SMILES sequences are guided by the synthetic-accessibility score and the synthetic-complexity score to ensure that polymer generation is close to that of synthetically accessible monomer-level structures. This framework addresses the challenge of generating novel polymer structures for laboratory researchers, thereby providing computational insights into polymer research.