Tea is one of the most popular drinks in the world. The rapid and accurate recognition of tea diseases is of great significance for taking targeted preventive measures. In this paper, an information entropy masked vision transformation (IEM-ViT) model was proposed for the rapid and accurate recognition of tea diseases. The information entropy weighting (IEW) method was used to calculate the IE of each segment of the image, so that the model could learn the maximum amount of knowledge and information more quickly and accurately. An asymmetric encoder–decoder architecture was used in the masked autoencoder (MAE), where the encoder operated on only a subset of visible patches and the decoder recovered the labeled masked patches, reconstructing the missing pixels for parameter sharing and data augmentation. The experimental results showed that the proposed IEM-ViT had an accuracy of 93.78% for recognizing the seven types of tea diseases. In comparison to the currently common image recognition algorithms including the ResNet18, VGG16, and VGG19, the recognition accuracy was improved by nearly 20%. Additionally, in comparison to the other six published tea disease recognition methods, the proposed IEM-ViT model could recognize more types of tea diseases and the accuracy was improved simultaneously.

Road blocking events refer to road traffic blocking caused by landslides, debris flow, snow disasters, rolling stones and other factors. To predict road blocking events, the limit gradient lifting model (XGBoost), random forest regression model (RF regression) and support-vector regression model (SVR) are used as the prediction meta-models, and then the meta-models are fused by a logical regression algorithm to construct a road blocking loss prediction fusion model based on ensemble learning. The actual road blocking event data are used to train the model. Using the same blocking location and similar blocking loss characteristics between adjacent points to fill in the missing value and conducting one-hot encoding for other short character sets with obvious category characteristics such as letters, numbers, and Chinese characters overcomes the problems of inherent data loss, error and time logic disorder in the blocking event data set. The test results show that the R² score based on the stacking fusion model reaches 0.91, which is 18% higher than RF and 11% and 5.8% higher than SVR and XGBoost, respectively, and the RMSE and MAE values are 0.1707 and 0.0341, respectively. Therefore, the proposed road blocking data preprocessing method and road blocking loss prediction fusion model can be used to predict the amount of blocking event loss.

Numerous studies have demonstrated that microRNAs (miRNAs) play crucial roles in the development and progression of various diseases, making the identification of miRNA-disease associations (MDAs) essential for understanding human disease etiology. While several computational models have been developed to predict MDAs, challenges persist—particularly the limited consideration of information interactions among multi-source similarities and the presence of “false-negative” associations in the original topology. To address these issues, we propose ISFNMDA, a model designed to infer potential MDAs by leveraging multi-view collaborative learning for feature extraction and optimizing association topology through graph structure momentum contrastive learning. Specifically, multi-source similarities of miRNAs and diseases are mapped into a unified feature space via encoders. The Pearson correlation coefficient is employed to derive pairwise constraints between nodes, facilitating information interactions and constructing interval-shared information constraints. Subsequently, an inference graph learner models the representations to generate an inferred graph topology. By maximizing mutual information between the inferred topology and the original “false-negative” associations through momentum contrastive learning, the model effectively reduces spurious correlations. The final comprehensive representations and optimized graph structure are then used to predict potential MDAs. Experimental results demonstrate that ISFNMDA outperforms existing methods, and case studies further validate its predictive capability. The complete code and related materials for ISFNMDA is available at https://github.com/WDNokl/ISFNMDA.

Deep learning foundation models are revolutionizing single-cell biology, yet learning holistic and discriminative representations from complex, high-dimensional data remains a central challenge. Although Transformer-based single-cell language models have shown significant progress, they typically rely on a single input-encoding scheme, a practice that results in the loss of critical gene expression information and hinders the effective learning of global cellular representations. To address these challenges, we introduce scDMC, an innovative single-cell Dual-stream Masked Contrastive pre-training framework designed to synergistically optimize information fidelity at both the gene and cellular levels. Pre-trained on only 2 million cells far fewer than the datasets used by mainstream models, scDMC sets a new state-of-the-art in multiple benchmark tasks, including cell annotation, clustering, and data integration. More importantly, we demonstrate that scDMC can uncover functional gene modules, infer cell-type-specific regulatory networks in a data-driven manner, and exhibits a high degree of biological interpretability.