Home Community This AI Paper Introduces a Groundbreaking Machine Learning Model for Efficient Hydrogen Combustion Prediction: Leveraging ‘Negative Design’ and Metadynamics in Reactive Chemistry

This AI Paper Introduces a Groundbreaking Machine Learning Model for Efficient Hydrogen Combustion Prediction: Leveraging ‘Negative Design’ and Metadynamics in Reactive Chemistry

0
This AI Paper Introduces a Groundbreaking Machine Learning Model for Efficient Hydrogen Combustion Prediction: Leveraging ‘Negative Design’ and Metadynamics in Reactive Chemistry

Potential energy surfaces (PESs) represent the connection between the positions of atoms or molecules and their associated potential energy. PESs are essential in understanding molecular behavior, chemical reactions, and material properties. They describe how the potential energy of a system changes because the positions of its constituent atoms or molecules vary. These surfaces are sometimes high-dimensional and complicated, making their accurate computation difficult, especially for big molecules or systems. 

The reliability of the machine learning ML model still heavily depends upon the range of the training data, especially for chemically reactive systems that must visit high-energy states when undergoing chemical transformations. ML models, by their nature, interpolate between known training data. Still, its extrapolation capability is restricted as predictions could be unreliable when molecules or their configurations are dissimilar to those within the training set. 

Formulating a balanced and diverse dataset for a given reactive system is difficult. It is not uncommon for the ML model to still suffer from an overfitting problem that may result in models with good accuracy on their original test set but are error-prone when applied to MD simulations, especially for gas phase chemical reactivity during which energy configurations are highly diverse.

Researchers on the University of California, Lawrence Berkeley National Laboratory, and Penn State University have built an lively learning AL workflow that expands the originally formulated Hydrogen combustion dataset by preparing collective variables (CVs) for the primary systematic sample. Their work reflects that a negative design data acquisition strategy is mandatory to create a more complete ML model of the PES. 

Following this lively learning strategy, they were in a position to achieve a final hydrogen combustion ML model that’s more diverse and balanced. The ML models get well accurate forces to proceed the trajectory without further retraining. They might predict the change within the transition state and response mechanism at finite temperature and pressure for hydrogen combustion.

The team has illustrated the lively learning approach on Rxn18 for instance during which the potential energy surface is projected onto two response coordinates, CN(O2-O5) and CN(O5-H4). The ML model performance was tracked by analyzing the unique data points derived from AIMD and normal modes calculations. They used longer metadynamics simulations for sampling because the lively learning rounds proceeded and errors decreased. 

They found metadynamics to be an efficient sampling tool for unstable structures, which helps the AL workflow discover holes within the PES landscape to tell the ML model through retraining with such data. Using metadynamics only as a sampling tool, the tricky CV selection step could be avoided by starting with reasonable or intuitive CVs. Their future work also includes analyzing alternate approaches like delta learning and dealing on more physical models like C-GeM.


Take a look at the Paper. All credit for this research goes to the researchers of this project. Also, don’t forget to affix our 33k+ ML SubReddit, 41k+ Facebook Community, Discord Channel, and Email Newsletter, where we share the newest AI research news, cool AI projects, and more.

In case you like our work, you’ll love our newsletter..


Arshad is an intern at MarktechPost. He’s currently pursuing his Int. MSc Physics from the Indian Institute of Technology Kharagpur. Understanding things to the elemental level results in recent discoveries which result in advancement in technology. He’s enthusiastic about understanding the character fundamentally with the assistance of tools like mathematical models, ML models and AI.


↗ Step by Step Tutorial on ‘Tips on how to Construct LLM Apps that may See Hear Speak’

LEAVE A REPLY

Please enter your comment!
Please enter your name here