Data-driven causal inference of process-structure relationships in nanocatalysis

Ting, Jonathan Y. C.; Barnard, Amanda S.

doi:10.1016/j.coche.2022.100818

Cited by 13 publications

(8 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In general, the reproducible workflow demonstrated here can be applied to any similar materials data set to make reliable model-agnostic predictions of how the structural characteristics and individual structure contribute to the prediction of functional properties. This provides a robust and reproducible approach to identify structure/property relationships that can be used as input to causal inference to predict why a structure gives rise to a property, , or inverse design to predict which structure to make to give a set of desirable properties. , …”

Section: Discussionmentioning

confidence: 99%

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

Barnard,

Fox

2023

Chem. Mater.

Self Cite

View full text Add to dashboard Cite

The application of machine learning (ML) to materials chemistry can accelerate the design process, and when coupled with a detailed explanation, can guide future research. Shapley value analysis is a complementary approach capable of providing a comprehensive analysis of the underlying reasons behind a structure/property relationship. In this study, we have used data sets of graphene oxide nanomaterials generated using electronic structure simulations to train ML models with outstanding accuracy, generalizability, and stability to predict the formation energy and the Fermi energy and applied Shapley value analysis to understand the results. Feature important profiles that rank the value of structural characteristics to each property confirmed that the underlying structure/property relationships are relatively simple and scientifically intuitive, even though the ML models need complex information to achieve high performance. We have also reported instance influence profiles that rank the value of each individual graphene oxide structure to the training process. Feature/instance interactions are also investigated to explain which structural characteristics make particular structures influential, revealing that the most influential structures typically have very high or very low concentrations of H or O. Since the range of concentrations is typically chosen by researchers based on domain knowledge at the outset, this highlights that extreme care should be taken when gathering training data as these decisions will have a very big impact on the final model once trained. In general, the reproducible workflow demonstrated here can be applied to any similar materials data set to make reliable model-agnostic predictions of how the structural characteristics and individual structures contribute to the prediction of functional properties.

show abstract

Section: Discussionmentioning

confidence: 99%

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

Barnard,

Fox

2023

Chem. Mater.

Self Cite

View full text Add to dashboard Cite

show abstract

“…On the other hand, causality approaches (causal inference, causal discovery) merged with ML methods can find out cause and effect relations among the features and target descriptors. Such kinds of approaches are in the nascent stage for catalysis , and have an enormous potential to explore the interesting cause and effect relations for catalytic descriptors and elemental features. Furthermore, new high-throughput experimental platforms are emerging using robotics to reduce human interventions and experimental errors. , All of these approaches are efficient and effective in terms of their respective applications.…”

Section: Challenges and Future Aspectsmentioning

confidence: 99%

A Route Map of Machine Learning Approaches in Heterogeneous CO₂ Reduction Reaction

et al. 2023

View full text Add to dashboard Cite

Machine learning (ML) with its indigenous predicting ability has been influential in the current scientific world and has enabled a paradigm shift in the field of CO2 reduction reaction (CO2RR). In this perspective, current research progress of ML approaches in heterogeneous electrocatalytic CO2RR has been demonstrated. The important findings related to the ML systems comprising features, output descriptors, and ML models have been summarized. Further, the opportunities and challenges in using the state-of-the-art ML methodologies along with the ways of circumventing those challenges are discussed. Finally, the interpretation of black box ML models and extensive usages of interpretable glass box and gray box models for CO2RR are encouraged for obtaining proper physical interpretations. The future directions on utilizing several such evolving ML methods to predict catalytic activity descriptors can help in a broader way to explore novel and efficient heterogeneous CO2RR and other similar catalytic reactions.

show abstract

“…Causal inference therefore complements conventional correlative ML applications as the results from correlational models provide insights that inform strategies ("doing the right things" by focusing on the important features) and the outcomes from causal models inform tactics ("doing things right" by learning the connection between the features). [28] The combination of multi-target regression and causal inference, and ideally the development of multi-target causal networks, simultaneously addresses the issues of trade-offs between related properties of multi-functional nanomaterials, and the translation of those relationships into a course of action.…”

Section: Introductionmentioning

confidence: 99%

Causal Paths Allowing Simultaneous Control of Multiple Nanoparticle Properties Using Multi‐Target Bayesian Inference

Ting

Barnard

2022

Advcd Theory and Sims

Self Cite

View full text Add to dashboard Cite

Machine learning can extract complex structure/property relationships but is often insufficient to explain how to control or tune the properties of materials, particularly when they are multi-functional. This study demonstrates the value of combining multi-target regression and multi-target causal graphs to address the need to simultaneously control multiple properties of nanomaterials, and the need to translate these relationships into actionable insights. Using nanodiamonds as an exemplar, recursive feature elimination is first used to identify nine structural features that allow simultaneous prediction of their electron charge transfer properties and thermochemical stability to high accuracy by an interpretable random forest regressor. A multi-target Bayesian network with domain knowledge incorporated via interactive learning using a hill-climbing algorithm then determines how these important structural features of nanodiamonds relate to their functional properties, proposing causal paths that can be used to inform experimental design.

show abstract

Data-driven causal inference of process-structure relationships in nanocatalysis

Cited by 13 publications

References 42 publications

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

A Route Map of Machine Learning Approaches in Heterogeneous CO₂ Reduction Reaction

Causal Paths Allowing Simultaneous Control of Multiple Nanoparticle Properties Using Multi‐Target Bayesian Inference

Contact Info

Product

Resources

About

Data-driven causal inference of process-structure relationships in nanocatalysis

Cited by 13 publications

References 42 publications

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

Importance of Structural Features and the Influence of Individual Structures of Graphene Oxide Using Shapley Value Analysis

A Route Map of Machine Learning Approaches in Heterogeneous CO2 Reduction Reaction

Causal Paths Allowing Simultaneous Control of Multiple Nanoparticle Properties Using Multi‐Target Bayesian Inference

Contact Info

Product

Resources

About

A Route Map of Machine Learning Approaches in Heterogeneous CO₂ Reduction Reaction