Nukunudompanich, Methawee

Loading...
Thumbnail Image
Preferred name
Nukunudompanich, Methawee
Alternative Name
Nukunudompanich, Meethawee
Main Affiliation
Email
methawee.nu@kmitl.ac.th

Settings

Now showing 1 - 3 of 3
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Machine learning of properties of lead-free perovskites with a neural network with additive kernel regression-based neuron activation functions
    (2024-07-01)
    Nukunudompanich, Methawee
    ;
    Yoon, Heejoo
    ;
    Hyojae, Lee
    ;
    Kameda, Keisuke
    ;
    Ihara, Manabu
    Machine learning (ML) of properties of perovskite materials, in particular of the bandgap of perovskites used in optoelectronic applications, has recently attracted increasing attention. Typically, off-the-shelf ML methods such as neural networks (NN), kernel methods or tree-based methods are used. We employ the recently proposed type of NN that uses additive Gaussian process regression to construct optimal neuron activation functions and avoids non-linear optimization to machine learn the band gap and heat of formation of lead-free inorganic halide double perovskites for solar cell applications. The method combines the high expressive power of an NN with the robustness of a linear regression. We show that better prediction quality can be obtained, in particular in the visible region relevant for most applications, compared to previous results using standard methods. Most important variables and the importance of coupling among features, in particular for bandgap prediction, can also be identified with the new method. Graphical abstract: (Figure presented.)
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    On the Sufficiency of a Single Hidden Layer in Feed-Forward Neural Networks Used for Machine Learning of Materials Properties
    (2025-03-01)
    Thant, Ye Min
    ;
    Manzhos, Sergei
    ;
    Ihara, Manabu
    ;
    Nukunudompanich, Methawee
    Feed-forward neural networks (NNs) are widely used for the machine learning of properties of materials and molecules from descriptors of their composition and structure (materials informatics) as well as in other physics and chemistry applications. Often, multilayer (so-called “deep”) NNs are used. Considering that universal approximator properties hold for single-hidden-layer NNs, we compare here the performance of single-hidden-layer NNs (SLNN) with that of multilayer NNs (MLNN), including those previously reported in different applications. We consider three representative cases: the prediction of the band gaps of two-dimensional materials, prediction of the reorganization energies of oligomers, and prediction of the formation energies of polyaromatic hydrocarbons. In all cases, results as good as or better than those obtained with an MLNN could be obtained with an SLNN, and with a much smaller number of neurons. As SLNNs offer a number of advantages (including ease of construction and use, more favorable scaling of the number of nonlinear parameters, and ease of the modulation of properties of the NN model by the choice of the neuron activation function), we hope that this work will entice researchers to have a closer look at when an MLNN is genuinely needed and when an SLNN could be sufficient.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Neural networks for neurocomputing circuits: A computational study of tolerance to noise and activation function non-uniformity when machine learning materials properties
    (2025-12-01)
    Thant, Ye min
    ;
    Nukunudompanich, Methawee
    ;
    Chueh, Chu Chen
    ;
    Ihara, Manabu
    ;
    Manzhos, Sergei
    Dedicated analog neurocomputing circuits are promising for high-throughput, low power consumption applications of machine learning (ML) and for applications where implementing a digital computer is unwieldy (remote locations; small, mobile, and autonomous devices, extreme conditions, etc.). Neural networks (NN) implemented in such circuits, however, must contend with circuit noise and the non-uniform shapes of the neuron activation function (NAF) due to the dispersion of performance characteristics of circuit elements (such as transistors or diodes implementing the neurons). We present a computational study of the impact of circuit noise and NAF inhomogeneity in regression problems as a function of NN architecture and training regimes. We focus on one application that requires high-throughput ML: materials informatics, using as representative problem ML of formation energies vs. lowest-energy isomer of peri-condensed hydrocarbons, formation energies and band gaps of double perovskites, and zero point vibrational energies of molecules from QM9 dataset. We show that in these applications, NNs generally possess low noise tolerance with the model accuracy rapidly degrading with noise level. Single-hidden layer NNs, and NNs with larger-than-optimal sizes are somewhat more noise-tolerant. Models that show less overfitting (not necessarily the lowest test set error) are more noise-tolerant. Importantly, we demonstrate that the effect of activation function inhomogeneity can be palliated by retraining the NN using practically realized shapes of NAFs.