Summary of Explainable Data-driven Modeling Of Adsorption Energy in Heterogeneous Catalysis, by Tirtha Vinchurkar et al.
Explainable Data-driven Modeling of Adsorption Energy in Heterogeneous Catalysis
by Tirtha Vinchurkar, Janghoon Ock, Amir Barati Farimani
First submitted to arxiv on: 30 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Chemical Physics (physics.chem-ph)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Medium Difficulty Summary: The paper explores the integration of machine learning (ML) and eXplainable AI (XAI) to bridge the gap between physics-based studies and data-driven methodologies in catalysis. The authors employ Post-hoc XAI analysis and Symbolic Regression to unravel the correlation between adsorption energy and properties of the adsorbate-catalyst system, leveraging the Open Catalyst Dataset (OC20). They use shallow ML techniques to predict adsorption energy and then apply post-hoc analysis to identify feature importance, inter-feature correlations, and the influence of various features on prediction. The study reveals that adsorbate properties have a greater influence than catalyst properties in their dataset. The top five features are adsorbate electronegativity, number of adsorbate atoms, catalyst electronegativity, effective coordination number, and sum of atomic numbers of the adsorbate molecule. There is a positive correlation between catalyst and adsorbate electronegativity with prediction of adsorption energy. Symbolic regression yields consistent results with SHAP analysis, demonstrating that the square of catalyst electronegativity is directly proportional to adsorption energy. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty Summary: This study uses machine learning (ML) and a special tool called eXplainable AI (XAI) to help design better catalysts. Catalysts are important in many industries, like chemical manufacturing. The researchers combined ML with XAI to understand how different properties of the adsorbate-catalyst system affect how well it works. They used a large dataset and found that the properties of the substance being adsorbed have a bigger impact than the properties of the catalyst itself. This study shows that by combining ML and XAI, we can better design catalysts for specific tasks. |
Keywords
* Artificial intelligence * Machine learning * Regression
