소재 AI 공유함
Menu
← AI in materials

AI trends · written by AI

SHAP interpretation reveals trade-offs in negative thermal expansion design for framework materials

AI editorPublished 2026.09.14

Interpretable machine learning and SHAP were used to analyze the relationship between structural descriptors and negative thermal expansion properties in framework materials. A trade-off between large negative thermal expansion and a wide temperature range was identified, yielding structure-property trends that can guide design. [1]

1/ AI utilization method: SHAP-based interpretable machine learning
2/ Applied task: analysis of the relationship between structural descriptors and the magnitude and temperature range of negative thermal expansion
3/ Reported result: identified a trade-off between large negative thermal expansion and a wide temperature window, and derived design trends
4/ The paper released on 2026-09-14 used interpretable machine learning and SHAP to analyze the relationship between structural descriptors and the magnitude and temperature range of negative thermal expansion(CNTE) in framework materials.
5/ As a result, it confirmed a trade-off between large negative thermal expansion and a wide temperature window, and presented structure-property trends that can be used as a design reference.
6/ From the standpoint of materials design, this result suggests that in complex structural systems it may be necessary to seek a balance between thermal expansion magnitude and the temperature operating range rather than maximizing both at once.
7/ SHAP-based interpretation helps identify which structural elements contribute strongly to the prediction.
8/ However, the reported trade-off depends on the dataset and the chosen structural descriptors, and it cannot be assumed that the same trend will hold unchanged for other framework families or different temperature ranges.
9/ The interpretive results are correlations and do not directly imply causation. [1]

How to apply this

Structure descriptors and property trade-offs can be found with interpretable models to determine design directions.

  1. Collect a dataset that includes both thermal expansion magnitude and temperature range, and calculate structural descriptors.
  2. Train an interpretable regression or classification model, then use SHAP to identify the main variables.
  3. Validate with independent data whether the trade-off between thermal expansion magnitude and temperature range is reproduced.
  4. Select candidate structures and verify the design trend with additional calculations or experiments.
Data needed
A materials dataset containing both structural descriptors and measured thermal expansion properties is needed.
Checking the result
Check prediction performance on an independent validation set, and see whether the SHAP ranking remains stable when bootstrapping or data splitting is changed.
What can go wrong
The importance ranking of an interpretable model is not the same as causality, and design guidance can become unstable when extrapolating to compositions outside the data range.

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

Related concepts and practice

이 이슈를 읽는 데 필요한 개념입니다. 용어마다 Colab에서 바로 돌려볼 수 있는 실습 노트북이 있습니다.

Sources

  1. Physical Review Materials - Accepted Paper: Interpretable machine learning reveals a structure-property trade-off between negative thermal expansion magnitude and temperature range in framework materials

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