The proliferation of spam in China has a negative impact on internet users’ experiences online. Existing methods for detecting spam are primarily based on machine learning. However, it has been discovered that these methods are susceptible to adversarial textual spam that has frequently been imperceptibly modified by spammers. Spammers continually modify their strategies to circumvent spam detection systems. Text with Chinese homophonic substitution may be easily understood by users according to its context. Currently, spammers widely use homophonic substitution to break down spam identification systems on the internet. To address these issues, we propose a Bidirectional Gated Recurrent Unit (BiGRU)–Text Convolutional Neural Network (TextCNN) hybrid model with joint embedding for detecting Chinese spam. Our model effectively uses phonetic information and combines the advantages of parameter sharing from TextCNN with long-term memory from BiGRU. The experimental results on real-world datasets show that our model resists homophone noise to some extent and outperforms mainstream deep learning models. We also demonstrate the generality of joint textual and phonetic embedding, which is applicable to other deep learning networks in Chinese spam detection tasks.

Precipitation nowcasting is very important for the sectors which critically depend on timely and accurate weather information. One of the challenges of precipitation nowcasting is radar echo extrapolation which predicts the radar echo images accurately. Nowadays, the methods of radar echo extrapolation are mostly based on ConvRNNs. Unfortunately, as lead time increases, these methods unavoidably suffer from the problem that high reflectivity values are underestimated. Therefore, we propose a forecast-refinement neural network based on DyConvGRU and U-Net to improve the predicting ability for high reflectivity during radar echo extrapolation. To improve the model’s ability to predict high reflectivities, dynamic convolution, and the forecast-refinement architecture are applied. And to obtain more realistic results, the WGAN’s training strategy is adopted to train the forecast module and the refinement module. Through experiments on a radar dataset from Shanghai, China, the results show that our proposed method obtains higher Probability of Detection (POD), Critical Success Index (CSI), Heidke Skill Score (HSS), and lower False Alarm Rate(FAR).

In this study, a deep learning method called Lightning-SN was developed and used for cloud-to-ground (CG) lightning identification. Based on artificial scenarios, this network model selects radar products that exhibit characteristic factors closely related to lightning. Advanced time of arrival and direction lightning positioning data were used as the labeling factors. The Lightning-SN model was constructed based on an encoder–decoder structure with 25 convolutional layers, five pooling layers, five upsampling layers, and a sigmoid activation function layer. Additionally, the maximum pooling index method was adopted in Lightning-SN to avoid characteristic boundary information loss in the pooling process. The gradient harmonizing mechanism was used as the loss function to improve the model performance. The evaluation results showed that the Lightning-SN improved the segmentation accuracy of the CG lightning location compared with the traditional threshold method, according to the 6-minute operating period of the current S-band Doppler radar, exhibiting a better performance in terms of lightning location identification based on high-resolution radar data. The model was applied to the Ningbo area of Zhejiang Province, China. It was applied to the lightning hazard prevention in the hazardous chemical park in Ningbo. The composite reflectivity and radial velocity were the two dominant factors, with a greater influence on the model performance than other factors.

Image-based virtual try-on provides customers with convenient online clothes selections by transferring garments onto a reference person. Despite the emergence of several solutions to generate photo-realistic images and adapt to complex poses, controlling clothing length remains a challenge. We argue that the clothing reconstruction did not consider clothing length information, which results in clothing length being uncontrollable in most virtual try-on methods. To overcome this limitation, a novel clothing-agnostic person representation is proposed, which eliminates clothing information and quantifies clothing length as a numerical value. A new segmentation generator is designed to predict try-on segmentation maps of any length conditioned on this representation. Moreover, we correct two inaccurate labels, which enables our model to generate a wider range of garment interactions in images, such as the top tucked into or worn over the bottom, as well as the top and bottom worn separately without intersecting. Extensive experiments demonstrate that our method achieves the goal of continuous clothing length control and generates photo-realistic images with fine details that outperform most baseline methods in terms of quantitative and qualitative metrics.

Accurate precipitation nowcasting is of great significance for flood prevention, agricultural production, and public safety. In recent years, spatiotemporal sequence models based on deep learning have been widely used for precipitation nowcasting and have achieved better prediction results than traditional methods. These models commonly use radar echo extrapolation and utilize the Z-R relationship between radar and rainfall to predict rainfall. However, radar echo data can be affected by various noises, and the Z-R correlation linking radar and rainfall encompasses several variables influenced by factors like terrain, climate, and seasonal variations. To solve this problem, we propose a dual-source attention dynamic neural network (DSADNet) for precipitation nowcasting, which is a network model that utilizes a fusion module to extract valid information from radar maps and rainfall maps, together with dynamic convolution and the attention mechanism, to directly predict future rainfall through encoding and decoding structure. We conducted experiments on a real dataset in Jiangsu, China, and the experimental results show that our model had better performance than the other examined models.