Polymer Graph Representation for Infinite Repeat Unit Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods struggle to accurately represent polymer structures in a computer-readable format, particularly for polymers with infinitely repeated repeat unit structures, leading to complex and incomplete representations that fail to capture their properties effectively.

Innovation Solution

A data structure and method that converts polymer chemical structures into graph data, incorporating node and edge attributes, including attach nodes and edges, to facilitate machine learning and property prediction using graph neural networks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If SMILES representation method is used to represent polymer structures, then the representation can be processed by computer devices, but it cannot properly represent polymers with infinitely repeated repeat unit structures

Engineering Contradiction:
Improveability to represent polymer structuresVSAvoidinability to represent infinite repetition
Core Design Contradiction:
Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent segments the polymer structure into a repeat unit structure that is repeated infinitely. Instead of representing the entire infinite polymer chain, it identifies and represents only the fundamental repeating unit, which captures the essential structure while avoiding the impossibility of representing infinite repetition.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses a copy-based approach by representing the repeat unit structure that can be mentally or computationally copied infinitely to reconstruct the full polymer structure. The SMILES representation includes symbols indicating repetition, allowing the finite representation to generate the infinite structure through copying operations.

Inventive Principle:
Principle #26Copying

2Loss of information

If BigSmILES or Hierarchical Descriptor methods are used to represent polymers, then more structural information can be captured, but the representation becomes more complex as the molecule size increases

Engineering Contradiction:
Improvecompleteness of polymer structure representationVSAvoidcomplexity of representation method
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent divides the polymer representation into a compact repeat unit structure with explicit repetition indicators. This segmentation allows the complex infinite structure to be represented by a simple, reusable unit, reducing the overall complexity compared to representing the entire polymer chain explicitly.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The repeat unit structure serves multiple functions: it represents the fundamental chemical structure, encodes the repetition pattern, and can be applied to any polymer with that repeat unit regardless of chain length. This universal representation method works for all polymers with repeating structures, simplifying the overall approach.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Device complexity

If fragmentary polymer representation methods are used, then processing becomes simpler, but the properties of polymers with infinitely repeated structures cannot be properly represented

Engineering Contradiction:
Improvesimplicity of processingVSAvoidaccuracy of polymer property representation
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent segments the polymer into a repeat unit that, while finite and simple to process, contains all the information needed to represent the infinite structure. This segmentation maintains processing simplicity while ensuring that polymer properties can be accurately predicted from the repeat unit characteristics.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The repeat unit structure is designed to be copied infinitely to reconstruct the full polymer. This copying mechanism ensures that the finite representation faithfully represents the infinite structure, maintaining reliability in property representation while keeping the actual data structure finite and manageable.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS20260088140A1Polymer Graph Neural Network and the Implementing Method Therefor
Publication Date: 2026.03.26 LG CHEM LTD
  • US20260088140A1 patent drawing
  • US20260088140A1 patent drawing
  • US20260088140A1 patent drawing

AI summary

A computer-implemented method for predicting property information of a polymer from a chemical structure of the polymer, a method for graphically representing the chemical structure of the polymer, and a method and system for producing property information of a polymer from graph information of the polymer by training an artificial neural network based on the chemical structure of the polymer using information prescribing an interconnection relationship between each atom of a plurality of atoms constituting a repeat unit structure of the polymer and an attach node to which the repeat unit structure is attached.