Xuexiang Li

Model compression can effectively reduce deep learning models’ computational and storage demands, facilitating their deployment and inference at the edge, especially for image analysis tasks. However, investigating practical compression algorithms for multi-task models is still in its infancy. We propose a multi-task deep learning pruning method for disentangled networks based on the lottery ticket hypothesis (MTDLP). First, the method introduces a task importance metric for assigning weights to each independent task. It implements a consistent pruning strategy to ensure that the performance of critical tasks is not affected when redundant parameters are eliminated. Second, the method uses a parameter decoupling module to untangle the intertwining of task parameters by separating the critical parameters of each task through a parameter mask. Finally, the pruned sparse network undergoes weighted random reinitialization and retraining to recover model accuracy and optimize overall performance. We evaluate the MTDLP method on mainstream multi-task image analysis datasets, including NYU-v2 and Cityscapes. The experimental results demonstrate that the MTDLP method outperforms mainstream methods, such as single-shot network pruning and iterative magnitude pruning, while maintaining a high degree of model sparsity, highlighting its potential for efficient multi-task image analysis in resource-constrained environments.

Medical knowledge graphs have emerged as essential tools for representing complex relationships among medical entities. However, existing methods for learning embeddings from medical knowledge graphs, such as DistMult, RotatE, ConvE, InteractE, JointE, and ConvKB, may not adequately capture the unique challenges posed by the domain, including the heterogeneity of medical entities, rich hierarchical structures, large-scale, high-dimensionality, and noisy and incomplete data. In this study, we propose an Adaptive Hierarchical Transformer with Memory (AHTM) model, coupled with a teacher–student model compression approach, to effectively address these challenges and learn embeddings from a rich medical knowledge dataset containing diverse entities and relationship sets. We evaluate the AHTM model on this newly constructed “Med-Dis” dataset and demonstrate its superiority over baseline methods. The AHTM model achieves substantial improvements in Mean Rank (MR) and Hits@10 values, with the highest MR value increasing by nearly 56% and Hits@10 increasing by 39%. Furthermore, we observe similar performance enhancements on the “FB15K-237” and “WN18RR” datasets. Our model compression approach, incorporating knowledge distillation and weight quantization, effectively reduces the model’s storage and computational requirements, making it suitable for resource-constrained environments. Overall, the proposed AHTM model and compression techniques offer a novel and effective solution for learning embeddings from medical knowledge graphs and enhancing our understanding of complex relationships among medical entities, while addressing the inadequacies of existing approaches.

In this study, we present a novel embedding model, named ResE, for predicting links in knowledge graphs. ResE employs depth-separable convolution and residual blocks, integrated with channel attention mechanisms. ResE surpasses previously published models, including the closely related TransE model, by achieving the satisfactory mean rank (MR) and the excellent Hits@10 scores on both WN18RR and FB15K-237 benchmarks. ResE is a promising model for knowledge graph completion tasks, with potential for further investigation and extension to new applications such as user-oriented relationship modeling. Although comparatively shallow compared to computer vision convolutional architectures, future work may explore deeper convolutional models. ResE exhibits remarkable performance and outperforms existing approaches, thus setting a new benchmark for knowledge graph completion. The outcomes of our study illustrate the effectiveness of incorporating depth-separable convolution and residual blocks, accompanied by channel attention mechanisms, in modeling knowledge graphs. These findings highlight ResE’s potential to push the boundaries of cutting-edge in this domain.

Wafer surface defect detection plays an important role in controlling product quality in semiconductor manufacturing, which has become a research hotspot in computer vision. However, the induction and summary of wafer defect detection methods in the existing review literature are not thorough enough and lack an objective analysis and evaluation of the advantages and disadvantages of various techniques, which is not conducive to the development of this research field. This paper systematically analyzes the research progress of domestic and foreign scholars in the field of wafer surface defect detection in recent years. Firstly, we introduce the classification of wafer surface defect patterns and their causes. According to the different methods of feature extraction, the current mainstream methods are divided into three categories: the methods based on image signal processing, the methods based on machine learning, and the methods based on deep learning. Moreover, the core ideas of representative algorithms are briefly introduced. Then, the innovations of each method are compared and analyzed, and their limitations are discussed. Finally, we summarize the problems and challenges in the current wafer surface defect detection task, the future research trends in this field, and the new research ideas.

In recent years, air pollutants have become an important issue in meteorological research and an indispensable part of air quality forecasting. To improve the accuracy of the Chinese Unified Atmospheric Chemistry Environment (CUACE) model’s air pollutant forecasts, this paper proposes a solution based on ensemble learning. Firstly, the forecast results of the CUACE model and the corresponding monitoring data are extracted. Then, using feature analysis, we screen the correction factors that affect air quality. The random forest algorithm, XGBoost algorithm, and GBDT algorithm are employed to correct the prediction results of PM_2.5, PM₁₀, and O₃. To further optimize the model, we introduce the grid search method. Finally, we compare and analyze the correction effect and determine the best correction model for the three air pollutants. This approach enhances the precision of the CUACE model’s forecast and improves our understanding of the factors that affect air quality. The experimental results show that the model has a better prediction error correction effect than the traditional machine learning statistical model. After the algorithm correction, the prediction accuracy of PM_2.5 and PM₁₀ is increased by 60%, and the prediction accuracy of O₃ is increased by 70%.