Air pollution is harmful to human health and restricts economic development, so predicting when and where air pollution will occur is a challenging and important issue, especially in fields of urban planning, factory production and human activities. In this paper, we propose a deep Spatio-Temporal Orthogonal Regularization Residual CNN (ST-OR-ResNet) for air prediction. Deep Convolutional Neural Network (CNN) is presented to capture the complex spatio-temporal relation of the dynamic biased meteorological data. Residual learning is designed to avoid unpredictable oscillations when training the network and verifying errors. For the issue of characteristic statistical migration and saddle point proliferation in deep network, the orthogonality regularizations are designed to stabilize the back-propagation errors, utilizing various advanced analytical tools such as restricted isometry property without extra hassle. We then benchmark their effects on public real-world datasets to demonstrate that ST-OR-ResNet has better predictive performance than the state-of-the-art methods.

Named Entity Recognition (NER) plays a crucial role in the field of Natural Language Processing, holding significant value in applications such as information extraction, knowledge graphs, and question–answering systems. However, Chinese NER faces challenges such as semantic complexity, uncertain entity boundaries, and nested structures. To address these issues, this study proposes an innovative approach, namely Multi-Granularity BERT Adapter and Efficient Global Pointer (MGBERT-Pointer). The semantic encoding layer adopts Multi-Granularity Adapter (MGA), while the decoding layer employs Efficient Global Pointer (EGP) network, ensuring collaborative functionality. The MGA, incorporating Character Adapter, Entity Adapter, and Lexicon Adapter through interactive mechanisms, are deeply integrated into the BERT base, significantly enhancing the model’s ability to handle complex contexts and ambiguities. The EGP, utilizing Rotary Position Embedding, resolves the issue of insufficient boundary information in traditional attention mechanisms, thereby improving the model’s understanding and recognition of nested entity structures. Experimental results on four public datasets demonstrate a significant enhancement in Chinese NER performance achieved by the MGBERT-Pointer model.

In order to utilize the distributed characteristic of sensors, distributed machine learning has become the mainstream approach, but the different computing capability of sensors and network delays greatly influence the accuracy and the convergence rate of the machine learning model. Our paper describes a reasonable parameter communication optimization strategy to balance the training overhead and the communication overhead. We extend the fault tolerance of iterative-convergent machine learning algorithms and propose the Dynamic Finite Fault Tolerance (DFFT). Based on the DFFT, we implement a parameter communication optimization strategy for distributed machine learning, named Dynamic Synchronous Parallel Strategy (DSP), which uses the performance monitoring model to dynamically adjust the parameter synchronization strategy between worker nodes and the Parameter Server (PS). This strategy makes full use of the computing power of each sensor, ensures the accuracy of the machine learning model, and avoids the situation that the model training is disturbed by any tasks unrelated to the sensors.

The explosion of genomic data provides new opportunities to improve the task of gene regulatory network reconstruction. Because of its inherent probability character, the Bayesian network is one of the most promising methods. However, excessive computation time and the requirements of a large number of biological samples reduce its effectiveness and application to gene regulatory network reconstruction. In this paper, Flooding-Pruning Hill-Climbing algorithm (FPHC) is proposed as a novel hybrid method based on Bayesian networks for gene regulatory networks reconstruction. On the basis of our previous work, we propose the concept of DPI Level based on data processing inequality (DPI) to better identify neighbors of each gene on the lack of enough biological samples. Then, we use the search-and-score approach to learn the final network structure in the restricted search space. We first analyze and validate the effectiveness of FPHC in theory. Then, extensive comparison experiments are carried out on known Bayesian networks and biological networks from the DREAM (Dialogue on Reverse Engineering Assessment and Methods) challenge. The results show that the FPHC algorithm, under recommended parameters, outperforms, on average, the original hill climbing and Max-Min Hill-Climbing (MMHC) methods with respect to the network structure and running time. In addition, our results show that FPHC is more suitable for gene regulatory network reconstruction with limited data.

Damaged image inpainting is one of the hottest research fields in computer image processing. The development of deep learning, especially Convolutional Neural Network (CNN), has significantly enhanced the effect of image inpainting. However, the direct connection between convolution layers may increase the risk of gradient disappearance or overfitting during training process. In addition, pixel artifacts or visual inconsistencies may occur if the damaged area is inpainted directly. To solve the above problems, we propose a novel Dense Gated Convolutional Network (DGCN) for generative image inpainting by modifying the gated convolutional network structure in this paper. Firstly, Holistically-nested edge detector (HED) is utilized to predict the edge information of the missing areas to assist the subsequent inpainting task to reduce the generation of artifacts. Then, dense connections are added to the generative network to reduce the network parameters while reducing the risk of instability in the training process. Finally, the experimental results on CelebA and Places2 datasets show that the proposed model achieves better inpainting results in terms of PSNR, SSIM and visual effects compared with other classical image inpainting models. DGCN has the common advantages of gated convolution and dense connection, which can reduce network parameters and improve the inpainting effect of the network.