Scholarlink

<p>Machine learning (ML) can advance autonomous experimentation in material synthesis by using predictive models to guide experiments, thus reducing the need for trial-and-error approaches. This study develops models to assist zeolitic imidazolate framework-8 (ZIF-8) synthesis by predicting outcomes for optimizing experiment conditions. We study the effect of hyperparameter tuning on XGBoost model performance and use genetic algorithm for optimal parameter selection. Additionally, the model provides interpretable predictions, which advance our understanding of the synthesis conditions that affect ZIF-8 morphology. These findings demonstrate how models can streamline materials synthesis and build the foundation for future autonomous experimentation.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data-driven approaches for autonomous experimentation: Insights into modeling and prediction dynamics of metal–organic frameworks synthesis

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Yuncheng Du,
  • Cristina Sanchez Cereceda,
  • Dongping Du

摘要

Machine learning (ML) can advance autonomous experimentation in material synthesis by using predictive models to guide experiments, thus reducing the need for trial-and-error approaches. This study develops models to assist zeolitic imidazolate framework-8 (ZIF-8) synthesis by predicting outcomes for optimizing experiment conditions. We study the effect of hyperparameter tuning on XGBoost model performance and use genetic algorithm for optimal parameter selection. Additionally, the model provides interpretable predictions, which advance our understanding of the synthesis conditions that affect ZIF-8 morphology. These findings demonstrate how models can streamline materials synthesis and build the foundation for future autonomous experimentation.

Graphical Abstract