Interpretable substructure-based screening identifies candidate polymer dielectrics with multiple properties

AI editorPublished 2026.09.28Posted 2026.09.29 12:01 KST

This materials informatics study evaluates candidate polymer dielectrics with interpretable screening at the substructure level and organizes the results into rational design rules. Because the approach emphasizes interpretability while narrowing the candidate pool, experimental synthesis and measurement must follow for verification. [1]

1/ AI 활용 방법: Interpretable substructure-based machine learning screening
2/ 적용 과제: Screening candidate polymer dielectrics with multiple properties
3/ 보고된 결과: Evaluated candidates and derived rational design rules
4/ This study is centered on screening candidate polymer dielectrics with interpretable substructure-based machine learning and organizing the results into design rules that can be used directly.
5/ The authors evaluated candidate structures at the partial-structure level and selected materials that satisfy multiple property criteria.
6/ The significance for materials science is that, in problems with a broad composition and structural space such as polymer dielectrics, the study goes beyond simple prediction to show, in an interpretable way, which structural elements contribute to candidate selection.
7/ This approach is useful for prioritizing synthesis and generating hypotheses.
8/ However, interpretable screening results do not directly guarantee experimental performance.
9/ Because substructure-based rules depend strongly on the range of the training data and the representation method, actual candidates must be confirmed with independent samples and experimental measurements. [1]

How to apply this

Score polymer dielectric candidates by substructure features and narrow down the synthesis targets to prioritize.

Decompose candidate polymer structures into substructures using a consistent rule set.
Organize labels and descriptors matched to the target properties, then split the data into training and validation sets.
Train a scoring model with interpretable machine learning and identify the important substructures.
Select top-ranked candidates for actual synthesis and measurement, then compare their rankings with the model output.

Data needed: Candidate polymer structural formulas, substructure annotations or molecular descriptors, and target property labels are required.
Checking the result: Check the agreement between predicted ranking and measured values in candidates not used for training and in newly synthesized samples.
What can go wrong: The most common mistake is overgeneralizing substructure rules beyond the data range and overtrusting them in unseen chemical space.

이 가이드는 원문의 방법을 바탕으로 AI가 제안한 적용 방향입니다. 원문에 적힌 내용이 아니며 같은 결과가 재현된다는 보장도 아닙니다.

웹 자료를 바탕으로 AI가 자동 요약한 글입니다. 연구의 세부 조건과 결과는 연결된 원문을 확인해주세요.

Interpretable substructure-based screening identifies candidate polymer dielectrics with multiple properties

How to apply this

Related concepts and practice

Sources