Search

Short text classification is a research focus for natural language processing (NLP), which is widely used in news classification, sentiment analysis, mail filtering and other fields. In recent years, deep learning techniques are applied to text classification and has made some progress. Different from ordinary text classification, short text has the problem of less vocabulary and feature sparsity, which raise higher request for text semantic feature representation. To address this issue, this paper propose a feature fusion framework based on the Bidirectional Encoder Representations from Transformers (BERT). In this hybrid method, BERT is used to train word vector representation. Convolutional neural network (CNN) capture static features. As a supplement, a bi-gated recurrent neural network (BiGRU) is adopted to capture contextual features. Furthermore, an attention mechanism is introduced to assign the weight of salient words. The experimental results confirmed that the proposed model significantly outperforms the other state-of-the-art baseline methods.

Traditional Tibetan medicine (TTM) is major component of traditional Chinese medicine (TCM). Over thousands of years, the Tibetan people have gradually established a comprehensive theoretical system for Tibetan medicine and obtained rich clinical experience in the snow-covered plateau. The history of TTM has developed rapidly, especially since the establishment of the People's Republic of China, and remarkable achievements have been made with the great support for ethnic medicines from the Communist Party of China and the central government. In this paper, we give an overview of the history, theoretical system, common treatment methods, current development, and prospects for the future development of TTM.

Campylobacter jejuni, one of the most important food-borne pathogens, develops resistance to macrolides by acquiring point mutations in the 23S rDNA gene (e.g. A2075G or A2074C) and in ribosomal protein L4 (e.g. G74D or G57D). In this study, a multiplex fluorescence real-time PCR assay, based on TaqMan minor groove binder (MGB) probes, was developed to identify C. jejunicontaining mutations associated with macrolide resistance. The VS1-MGB probe was designed based on the VS1gene and was used to identify C. jejuni. The 23S rDNA-MGB probe was designed to distinguish macrolide resistance mutations in 23S rDNA, whilst 57D-MGB and 74D-MGB were designed to detect resistance mutations in ribosomal protein L4. The specificity and accuracy of the established method were 100% identical to the classic biochemical method, mapAPCR, minimum inhibitory concentration (MIC) determination and DNA sequencing. The linear detection limitation of the method was 0.03ng/reaction genomic DNA and 3 CFU/reaction bacterial colonies. In conclusion, this study provides a novel method that can specifically identify C. jejuniand accurately detect the gene mutations associated with macrolide resistance.

Samples of DLC thin films with different Cu concentration were prepared by using a Cu target and different number of graphite target，or adding a grid before the Cu target. The results show that the hardness and stress decreases with an increase of Cu concentration (below 6.3 at.%). When Cu concentration increases from 6.3 at.% to 17.8 at.%, the hardness and stress keep almost constant, and the friction coefficient increases gradually and slowly, as well as the wear loss.

Straight river is generally regarded as one of the typical river patterns in conventional classifications in terms of their channel plain landforms. However, very few straight patterns were found to be distributed in wider spatial and temporal spans in the self-adjusted fluvial rivers. Thus, the questions occur such as that is it possible for a channel takes on a stable straight pattern? What are the main factors controlling the processes of the river pattern formation and transformation from a straight to other patterns? Various theories and hypotheses including geomorphic threshold hypothesis, the extreme hypothesis on energy dissipation rate, the stability theory, etc. have been developed to explain the aforementioned questions, but none of them is sound for the explanation to the straight-river formation. From the modern fluvial plain patterns, the straight patterns are not as stable as other typical patterns which occurred in nature; from the historic records of the river sedimentation, no apparent evidence was found to support the stable straight river evolution. Based on the analysis of existing theories, observations, evolvement processes of the channel patterns in the experimental results, this paper concluded that the straight pattern should not be included as one of the typical patterns that are self-formed and developed. This study is of importance to understanding of the river pattern formation and transformation.