Optimizing the Prediction of Adsorption in Metal-Organic Frameworks Leveraging Q-Learning

Etinosa Osaro,Yamil Colón

DOI: https://doi.org/10.26434/chemrxiv-2024-3chlm

2024-04-04

Abstract:The application of machine learning (ML) techniques in materials science has revolutionized the pace and scope of materials research and design. In the case of metal-organic frameworks (MOFs), a promising class of materials due to their tunable properties and versatile applications in gas adsorption and separation, ML has helped survey the vast material space. This study explores the integration of reinforcement learning (RL), specifically Q-learning, with Gaussian processes (GPs) for predictive modeling of adsorption in MOFs. We demonstrate the effectiveness of the RL-driven framework in guiding the selection of training data points and optimizing predictive model performance for methane and carbon dioxide adsorption, using two different reward metrics. Our results highlight the adaptability and versatility of RL in navigating the adsorption predictions in MOFs, with the integration of GPs enhancing the robustness and reliability of predictive modeling.

Chemistry

What problem does this paper attempt to address?

The paper attempts to address the problem of optimizing adsorption prediction in metal-organic frameworks (MOFs). Specifically, the authors explore how to combine reinforcement learning (RL), particularly Q-learning, with Gaussian processes (GPs) to improve the prediction performance of methane (CH₄) and carbon dioxide (CO₂) adsorption in MOFs. Through this approach, the study aims to guide the selection of training data points, thereby optimizing the performance of the prediction model.

Key points of the paper include:
1. **Problem Background**: MOFs are highly regarded for their tunable properties and wide applications in gas adsorption, separation, and other fields. However, due to the vast MOF database, more advanced computational methods are needed to efficiently screen these materials.
2. **Method Innovation**: The study introduces Q-learning as a method for dynamically selecting the training dataset, which differs from traditional static selection methods. Q-learning can gradually optimize the selection of training data points through trial-and-error learning, thereby enhancing the accuracy of the prediction model.
3. **Technical Details**: By combining Gaussian Process Regression (GPR), the Q-learning algorithm can select new data points that most improve prediction accuracy based on the current state (i.e., the existing training dataset) in each iteration. The reward mechanism guides the learning process of the algorithm based on changes in prediction error (such as MRE or R² score).
4. **Experimental Validation**: The study first conducted experiments on two MOFs (Cu-BTC and IRMOF-1) to validate the effectiveness of the Q-learning framework. Subsequently, the method was extended to 8671 MOFs in the CoRE MOF database to further evaluate its generalization ability.

Overall, the paper proposes a new method to optimize the prediction of adsorption behavior in MOFs by combining Q-learning and GPs, aiming to improve the accuracy and reliability of the prediction model.