In materials science, the relationship between the material internal structure and its associated macroscale properties can be used to guide the design of materials. In this study, we constructed an interpretative machine learning model to capture the structure-property relationship and predict the solid solubility in binary alloy systems. To do this, we used a dataset containing about 1,843 binary alloys and their respective experimental values of solid solubility. We designed a common function to represent the relationship between individual descriptor and solid solubility, and a deep neural network to integrate the multiple functions. The resulting model can correctly predict the solid solubility value within less than 6% error. What's more, based on this model, it's feasible to analyze the effect of structures on target property.

Rapid antimicrobial resistance (AMR) prediction from MALDI-TOF mass spectrometry (MS) remains challenging, particularly when training artificial intelligence (AI) models under small-sample constraints. Performance is often hampered by the high dimensionality of spectral data and the subtle nature of resistance-related signals: full-spectrum approaches risk overfitting to high-dimensional noise, whereas peak-selection strategies risk discarding structurally informative, low-intensity signals. Here, we propose ReShuffle-MS, a region-guided data augmentation framework for MS data. Each spectrum is partitioned into a Main Discriminative Region (MDR) and a Peripheral Peak Region (PPR). By recombining signals within the PPR across samples of the same class while keeping the MDR intact, ReShuffle-MS generates structure-preserving augmented samples. On a clinical dataset for Escherichia coli (E. coli) levofloxacin resistance prediction, ReShuffle-MS delivered significant and consistent performance gains. It improved the average accuracy of classical machine learning models by 3.7% and enabled a one-dimensional convolutional neural network (CNN) to achieve 83.25% accuracy and 97.28% recall. Visualization using Grad-CAM revealed a shift from sparse, peak-dependent attention toward broader and more meaningful spectral patterns. Validation on the external DRIAMS-C dataset for ceftriaxone resistance further demonstrated that the method generalizes to a distinct laboratory setting and a different antibiotic target. These findings suggest that ReShuffle-MS can enhance the robustness and clinical utility of AI-based AMR prediction from routinely acquired MALDI-TOF spectra.

Copolymers' diverse properties are influenced by monomer type, stoichiometry, and chain architecture, making prediction challenging. Traditional machine learning models struggle to capture these complexities. This study introduces Co-MPNN, a novel graph-based model integrating stoichiometric ratios and chain architectures into a message-passing neural network (MPNN). Using a copolymer dataset from DFT calculations, Co-MPNN outperforms baseline models like MPNN and Random Forest in predicting electron affinity (EA) and ionization potential (IP). Sensitivity analysis confirms Co-MPNN's robustness with small training sets, while latent space visualization highlights its capacity to prioritize key features, enhancing interpretability. These results demonstrate Co-MPNN's potential for accurate and interpretable copolymer property prediction, advancing copolymer design and optimization.