2024 · Machine Learning: Science and Technology

Multimodal neural network-based predictive modeling of nanoparticle properties from pure compounds

Can Polat , Mustafa Kurban , Hasan Kurban

Research highlights

Pure2DopeNet: learning doped properties from pure compounds

Predict nanoparticle properties using image and text representations derived from a pure-compound simulation.

Why we investigated it

Repeating simulations across doping levels can make candidate exploration expensive.

What the work contributes

Pure2DopeNet combines modalities to predict properties of carbon-doped TiO2 and sulfur-doped ZnO.

What the evidence shows

The reported comparisons show improved prediction over the tested image-model baselines and agreement with DFTB targets in the studied systems.

Build on this work

Use the paper's pure-to-doped formulation to test how much simulation work a surrogate can replace within a specified chemistry.

  • Multimodal learning
  • Nanoparticle properties
  • Doping
  • DFTB surrogates
Abstract
Abstract Simulating complex and large materials is a challenging task that requires extensive domain knowledge and computational expertise. This study introduces Pure2DopeNet, an innovative multimodal neural network that tackles these challenges by integrating image and text data to accurately predict the physical properties of doped compounds, specifically Carbon (C)-doped TiO2 and Sulfur (S)-doped ZnO nanoparticles. The model achieves quantum mechanical level accuracy, comparable to density functional tight binding (DFTB), across various doping levels, demonstrating its capability to determine the properties from a single simulation of the pure compound. Pure2DopeNet outperforms traditional deep learning architectures such as ResNet, ViT, and CoAtNet, delivering superior accuracy, faster performance, and reduced dependence on domain expertise. This approach highlights the potential of multimodal machine learning to revolutionize materials science by making high-fidelity simulations more accessible and efficient, opening paving the way for material discovery and the exploration of novel properties.
Cite
@article{polat2024multimodal,
  title = {Multimodal neural network-based predictive modeling of nanoparticle properties from pure compounds},
  author = {Can Polat and Mustafa Kurban and Hasan Kurban},
  year = 2024,
  journal = {Machine Learning: Science and Technology},
  doi = {10.1088/2632-2153/ad9708},
}