An active area of MGI (Materials Genome Initiative)/MGE (Materials Genome Engineering) is to accelerate the development of new materials by means of active learning and “digital trial‐error” using a prediction model of material property. Machine learning methods have widely been employed for predicting crystalline materials properties with crystal graph neural networks (CGNN). The prediction accuracy of the state‐of‐the‐art (SOTA) CGNN models based on big models and big data is generally higher. However, for the development of some classes of materials, the datasets obtained by experiments are usually lacking due to costly experiments and measurement costs. The lack of datasets will impact the accuracy of CGNN models and may result in overfitting during training models. This paper proposes a simplified crystal graph convolutional neural network (S‐CGCNN) which possesses higher prediction accuracy while reducing the vast amount of train datasets and computation costs. The S‐CGCNN model has successfully predicted properties of crystalline materials, such as piezoelectric materials and dielectric materials, and increased the prediction accuracy up to 12%–20% than existing SOTA CGNN models. Furthermore, the distribution map between properties and compositions of materials has been built to screen the latent space of candidate materials efficiently by principal component analysis.

The development of novel refractory high‐entropy alloys (RHEAs) holds significant promise for advanced applications due to their exceptional properties. However, identifying optimal compositions of RHEAs within the vast alloy design space to meet specific property requirements remains a formidable challenge. In this study, we present an integrated machine learning (ML) framework to address this challenge, combining predictive models for material properties, a fingerprint map of composition distribution, a guided multiobjective search strategy, and a particle swarm optimizer to enable targeted exploration of promising RHEAs compositions. Using this approach, we successfully discovered several new RHEAs with outstanding mechanical performance, including Nb _0.189 Ti _0.203 V _0.203 Mo _0.206 Zr _0.197 , Nb _0.204 Ti ₀₁₉ V _0.207 Mo _0.198 Zr _0.198 , Nb _0.174 Ti _0.19 V _0.251 Mo _0.201 Zr _0.181 , Nb _0.242 Ti _0.252 To _0.001 V _0.039 Mo _0.209 Zr _0.254 , and Nb _0.164 Ta _0.155 Ti _0.186 V _0.008 W _0.153 Mo _0.001 Hf _0.168 Zr _0.16 . These alloys exhibit remarkable yield strengths ranging from 1580 to 1740 MPa and fracture strains between 23% and 27%. The integrated ML models make it possible to rapidly optimize multiple properties during other materials designing, thus overcoming the common problems of limited data and a vast composition space in complex materials systems, paving the way for efficient design of advanced materials tailored to diverse application requirements.

The Material Genome Initiative has been driven by high-throughput calculations, experiments, characterizations, and machine learning, which has accelerated the efficiency of the discovery of novel materials. However, the precise quantification of the material microstructure features and the construction of microstructure–property models are still challenging in optimizing the performance of materials. In this study, we proposed a new model based on machine learning to enhance the power of the data augmentation of the micrographs and construct a microstructure–property linkage for cast austenitic steels. The developed model consists of two modules: the data augmentation module and microstructure–property linkage module. The data augmentation module used a multi-layer convolution neural network architecture with diverse size filter to extract the microstructure features from irregular micrographs and generate new augmented microstructure images. The microstructure–property linkage module used a modified VGG model to establish the relationship between the microstructure and material property. Taking cast austenitic stainless steels after solution treating in different temperatures as an example, the results showed that the prediction accuracy of the developed machine learning model had been improved. The coefficient R² of the model was 0.965, and the medians were only ±2 J different with the measured impact toughness.