Text semantic similarity is a crucial research area in Natural Language Processing (NLP). However, traditional methods for calculating the similarity of short Chinese texts often fall short in accuracy and other cruci...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
Text semantic similarity is a crucial research area in Natural Language Processing (NLP). However, traditional methods for calculating the similarity of short Chinese texts often fall short in accuracy and other crucial aspects. The Bidirectional Long Short-Term Memory Network (BiLSTM) has demonstrated remarkable performance in computing the similarity of short Chinese texts by effectively capturing long-range dependencies and semantic information. Additionally, the multi-head attention mechanism considers the interaction between different text locations and semantic information, enhancing the model's representative capacity. Building upon this foundation, our paper proposes an enhanced model known as SCTbilstmAt-tRdrop (SCTAR). This model is constructed based on a multi-layer BiLSTM architecture and incorporates an SE-gated convolutional module and a convolutional multi-head attention-aware model. We conducted extensive evaluations using two Chinese short text datasets, Chinese-SNLI and CCKS2018_Task3. The experimental results unequivocally demonstrate that our SCTAR model surpasses other common methods in terms of accuracy, precision, recall, and F1 scores when tasked with computing the similarity of Chinese short texts.
Promoter components play a critical role in the regulation of gene expression, directly determining the expression intensity of downstream target genes. Opting for high-quality promoters is essential for synthetic bio...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
Promoter components play a critical role in the regulation of gene expression, directly determining the expression intensity of downstream target genes. Opting for high-quality promoters is essential for synthetic biology. In the existing literature, the method of designing high-quality promoters using Generative Adversarial Networks (GANs) is constrained by training difficulties, mode collapse, implicit generation, and information loss in deep networks after vector representation of promoters. Diffusion models are more easily trainable, possess elegant mathematical explanations, and can directly model the target distribution, potentially overcoming the issues faced by GANs mentioned above. In this paper, we propose a new method for promoter design, PromoterDiff, based on the diffusion model. Specifically, we adjusted the convolutional network structure of the diffusion model and introduced a bridging structure to adapt to the diffusion and reconstruction steps of the diffusion model when dealing with DNA sequence tensors containing a large number of zero elements. We used natural promoters of Escherichia coli as the training set to train the model, resulting in the successful design of 14,080 entirely new promoters. Through the analysis of the motifs, k-mer frequencies, and motif spacing constraints of these promoters, we confirmed that the diffusion model can capture the characteristics of natural promoters, and the quality of the generated promoters surpasses existing models. Through biological experiments, it has been confirmed that 83% of the promoter sequences designed by PromoterDiff surpass the activity of natural promoters, exceeding the current leading model by 13%. This underscores the immense potential of applying the diffusion model to de novo promoter design. This paper also provides a clear mathematical representation of the de novo promoter design task. According to our survey, this is the first time the diffusion model has been applied in the field
Few-shot medical image segmentation is a prominent area of collaborative medical technology research to support healthcare and homecare, but often grapples with challenges, such as the problem of local information los...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
Few-shot medical image segmentation is a prominent area of collaborative medical technology research to support healthcare and homecare, but often grapples with challenges, such as the problem of local information loss in most of the current prototype-based methods. To mitigate this concern, we propose a novel Multi-Level Feature-Guided Network (MLFGNet) for few-shot medical image segmentation, which aims to enhance segmentation performance by extracting richer information from multi-level features. Specifically, A Multi-Feature Processing Module is proposed that utilizes hybrid attention mechanisms working together to efficiently learn more useful features. Furthermore, we design a Prediction Aggregation Module to aggregate multiple segmentation prediction maps by assigning appropriate weights, thereby effectively improving segmentation accuracy. Extensive experiments prove that our method exhibits favorable performance with respect to the state-of-the-art methods on two public datasets, including Abdominal-CT and Abdominal-MRI datasets.
Nuclei detection and segmentation are indispensable prerequisites in digital pathology research, whereas the precise segmentation of nuclei by domain experts relies heavily on global spatial information and the inter-...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
Nuclei detection and segmentation are indispensable prerequisites in digital pathology research, whereas the precise segmentation of nuclei by domain experts relies heavily on global spatial information and the inter-nuclei correlation. However, previous automatic nuclei segmentation works are mostly built on convolutional neural networks, which are unable to capture long-range global context with their inherent convolutional operations. Additionally, window-based design in few transformerbased approaches limits remote token interactions. The present study introduces a novel Multi-Local Perception (MLP) network, MLPSeg, which is proposed to address the aforementioned challenging issues. Specifically, the parallel computation of depthwise separable convolution and local window attention is designed to extract local information. Then, the parallel module of local horizontal attention and local vertical attention is designed to establish the global dependency. Moreover, to model the crossscale dependencies and narrow the contextual semantic gap, the Context Cross Attention (CCA) is introduced for optimising skip connections. A tri-decoder structure is adopted to generate nuclei instance masks, normal edge masks and clustered edge masks. The superior performance of MLPSeg for nuclei segmentation is demonstrated across two datasets with different modalities, resulting in a 2.29% - 5.82% improvement compared to the state-of-the-art methods.
According to recent research, the hybrid workspace should be recognized as a distinct "third space'' alongside the fully physical and fully remote workspace. However, the specific attributes of this third...
详细信息
This paper advances the understanding of the digital and cultural significance of the online community colloquially known as the "black manosphere."Through the use of Critical Technocultural Discourse Analys...
详细信息
Developing regions lack personalizable mobile applications with cultural context for CWA to aid in communication with parents and teachers. This led us to design and develop an application to address this underexplore...
详细信息
As mental health influences many aspects of the college student's life, including academic performance and interpersonal relationships, it has become an area of interest at many American universities. Previous stu...
详细信息
暂无评论