arXiv subjects

Masato Sumita

Publications and source records attributed to Masato Sumita.

4 recordsLinked to original sources

arXiv2024

Molecule Graph Networks with Many-body Equivariant Interactions

Message passing neural networks have demonstrated significant efficacy in predicting molecular interactions. Introducing equivariant vectorial representations augments expressivity by capturing geometric data symmetries, thereby improving model accuracy. However, two-body bond vectors in opposition may cancel each other out during message passing, leading to the loss of directional information on their shared node. In this study, we develop Equivariant N-body Interaction Networks (ENINet) that explicitly integrates l = 1 equivariant many-body interactions to enhance directional symmetric information in the message passing scheme. We provided a mathematical analysis demonstrating the necessity of incorporating many-body equivariant interactions and generalized the formulation to $N$-body interactions. Experiments indicate that integrating many-body equivariant representations enhances prediction accuracy across diverse scalar and tensorial quantum chemical properties.

cs.LG↗

arXiv2022

Automatic Differentiation for the Direct Minimization Approach to the Hartree-Fock Method

Automatic differentiation has become an important tool for optimization problems in computational science, and it has been applied to the Hartree-Fock method. Although the reverse-mode automatic differentiation is more efficient than the forward-mode, eigenvalue calculation in the self-consistent field method has impeded the use of the reverse-mode automatic differentiation. Here, we propose a method to directly minimize Hartree-Fock energy under the orthonormality constraint of the molecular orbitals using reverse-mode automatic differentiation by avoiding eigenvalue calculation. According to our validation, the proposed method was more stable than the conventional self-consistent field method and achieved comparable accuracy.

physics.chem-ph↗

arXiv2019

Optimization of heterogeneous ternary Li3PO4-Li3BO3-Li2SO4 mixture for Li-ion conductivity by machine learning

Mixing heterogeneous Li-ion conductive materials is one of potential ways to enhance the Li-ion conductivity more than that of the parent materials. However, the development of the mixtures had not exhibited significant progress because it is a formidable task to cover the vast possible composition of the parent materials using traditional ways. Here, we introduce a fashion based on machine learning to optimize the composition ratio of ternary Li3PO4-Li3BO3-Li2SO4 mixture for its Li-ion conductivity. According to our results, the optimum composition of the ternary mixture system is 25:14:61 (Li3PO4: Li3BO3: Li2SO4 in mol%), whose Li-ion conductivity is measured as 4.9 x 10E-4 S/cm at 300 {\deg}C. Our X-ray structure analysis indicates that Li-ion conductivity in the mixing systems is enhanced by virtue of the coexistence of two or more phases. Although the mechanism enhancing Li-ion conductivity is not simple, our results demonstrate the effectiveness of machine learning for the development of materials.

cond-mat.mtrl-sci↗

arXiv2019

Leveraging Legacy Data to Accelerate Materials Design via Preference Learning

Machine learning applications in materials science are often hampered by shortage of experimental data. Integration with legacy data from past experiments is a viable way to solve the problem, but complex calibration is often necessary to use the data obtained under different conditions. In this paper, we present a novel calibration-free strategy to enhance the performance of Bayesian optimization with preference learning. The entire learning process is solely based on pairwise comparison of quantities (i.e., higher or lower) in the same dataset, and experimental design can be done without comparing quantities in different datasets. We demonstrate that Bayesian optimization is significantly enhanced via addition of legacy data for organic molecules and inorganic solid-state materials.

cond-mat.mtrl-sci↗