Using deep neural network with small dataset to predict material defects

TLDR

Deep neural networks achieve state‑of‑the‑art performance in many domains, yet when trained on small material science datasets they underperform compared to traditional machine learning, limiting their widespread use. This study aims to predict solidification defects using a DNN regression model trained on a small dataset of 487 samples. The authors employed a pre‑trained DNN that was fine‑tuned on the 487‑point dataset to perform regression of defect outcomes. The fine‑tuned pre‑trained DNN achieved superior generalization compared to shallow neural networks, support vector machines, and conventionally trained DNNs, producing a high‑accuracy map of defect likelihood across high‑dimensional chemistry and processing parameters, and demonstrating that small‑dataset DNNs can be viable when large datasets are unavailable.

Abstract

Deep neural network (DNN) exhibits state-of-the-art performance in many fields including microstructure recognition where big dataset is used in training. However, DNN trained by conventional methods with small datasets commonly shows worse performance than traditional machine learning methods, e.g. shallow neural network and support vector machine. This inherent limitation prevented the wide adoption of DNN in material study because collecting and assembling big dataset in material science is a challenge. In this study, we attempted to predict solidification defects by DNN regression with a small dataset that contains 487 data points. It is found that a pre-trained and fine-tuned DNN shows better generalization performance over shallow neural network, support vector machine, and DNN trained by conventional methods. The trained DNN transforms scattered experimental data points into a map of high accuracy in high-dimensional chemistry and processing parameters space. Though DNN with big datasets is the optimal solution, DNN with small datasets and pre-training can be a reasonable choice when big datasets are unavailable in material study.