The latent variable prior of the variational autoencoder (VAE) often utilizes a standard Gaussian distribution because of the convenience in calculation, but has an underfitting problem. This paper proposes a variational autoencoder with optimizing Gaussian mixture model priors. This method utilizes a Gaussian mixture model to construct prior distribution, and utilizes the Kullback-Leibler (KL) distance between posterior and prior distribution to implement an iterative optimization of the prior distribution based on the data. The greedy algorithm is used to solve the KL distance for defining the approximate variational lower bound solution of the loss function, and for realizing the VAE with optimizing Gaussian mixture model priors. Compared with the standard VAE method, the proposed method obtains state-of-the-art results on MNIST, Omniglot, and Frey Face datasets, which shows that the VAE with optimizing Gaussian mixture model priors can learn a better model.

Urban water plays a significant role in the urban ecosystem, but urban water extraction is still a challenging task in automatic interpretation of synthetic aperture radar (SAR) images. The influence of radar shadows and strong scatters in urban areas may lead to misclassification in urban water extraction. Nevertheless, the local features captured by convolutional layers in Convolutional Neural Networks (CNNs) are generally redundant and cannot make effective use of global information to guide the prediction of water pixels. To effectively emphasize the identifiable water characteristics and fully exploit the global information of SAR images, a modified Unet based on hybrid attention mechanism is proposed to improve the performance of urban water extraction in this paper. Considering the feature extraction ability and the global modeling capability in SAR image segmentation, the Channel and Spatial Attention Module (CSAM) and the Multi-head Self-Attention Block (MSAB) are both introduced into the proposed Hybrid Attention Unet (HA-Unet). In this work, Resnet50 is adopted as the backbone of HA-Unet to extract multi-level features of SAR images. During the feature extraction process, CSAM based on local attention is adopted to enhance the meaningful water features and ignore unnecessary features adaptively in feature maps of two shallow layers. In the last two layers of the backbone, MSAB is introduced to capture the global information of SAR images to generate global attention. In addition, two global attention maps generated by MSAB are aggregated together to reconstruct the spatial feature relationship of SAR images from high-resolution feature maps. The experimental results on Sentinel-1A SAR images show that the proposed urban water extraction method has a strong ability to extract water bodies in the complex urban areas. The ablation experiment and visualization results vividly indicate that both CSAM and MSAB contribute significantly to extracting urban water accurately and effectively.

Automatic text classification is a fundamental task in the field of natural language processing and it can help users select vital information from massive text resources. To better represent the semantic meaning of a text, and to solve the problem that traditional methods need to extract features manually, we use TF-IDF algorithm to calculate the weight of each word in a text, then weight the word vectors by TF-IDF value. This method will generate text vectors, which have clearer semantic meanings. Then we input the text vector matrix into Convolution Neural Network (CNN), so that the CNN will automatically extract text features. Through extensive experiments conducted on two data sets, experiments demonstrate that our approach can effectively improve the accuracy of classification, and the classification accuracy of the two data sets are 96.28% and 96.97% respectively.

Guessing random additive noise decoding (GRAND) is a noise-centric universal algorithm that is suitable for linear block codes. Using the guessing decoding independence of GRAND-Markov order (GRAND-MO), in this paper, a serial-storing and parallel-crossing configured high throughput GRAND-MO decoding scheme is proposed. Subsequently, the key permutation generation module, the noise error patterns (NEPs) generation module and the guessing decoding module, are each elucidated through a comprehensive step-by-step approach. The NEPs generation module outputs patterns solely based on burst parameters. For the targeted cyclic redundancy check (CRC) codes, the error detection capability is jointly integrated with GRAND-MO for error correction. Compared to conventional linear block codes, simulation results show the error correction capability of CRC is significantly improved. Given a packet length of 128 bits and a clock frequency of 866 MHz, field programmable gate array (FPGA) implementation shows that the NEPs generation throughput can reach 14.6 Gbps under the 128 parallelism assumption. The throughput can be flexibly adjusted to satisfy different throughput requirements at the cost of increased hardware overhead. This work serves as a practical reference for future research in GRAND-MO.