It is a key issue to reasonably represent human travel and social contact in epidemic models. Various measures were applied to develop the models of human mobility and contact in a long range or a short range, such as...
详细信息
Teaching early warning is of great significance for avoiding teaching risks and continuously improving teaching quality. However, none of existing approaches assess the degree of course goals attainment and teacher...
详细信息
ISBN:
(数字)9798350349184
ISBN:
(纸本)9798350349191
Teaching early warning is of great significance for avoiding teaching risks and continuously improving teaching quality. However, none of existing approaches assess the degree of course goals attainment and teacher's teaching quality from the perspective of cognitive diagnosis. This poses a challenge in providing accurate teaching early warning. This paper proposed an early warning approach for teachers based on cognitive diagnosis and long short-term memory (LSTM). First, this approach accurately evaluates students' cognitive status on knowledge concepts using a cognitive diagnosis model to assess their knowledge understanding degree and knowledge application ability. Second, the cognitive status on knowledge concepts is utilized to assess students' attainment degree of course goals and teachers' teaching quality. Third, the teachers' teaching quality is predicted in the future by using the LSTM network to mine students' learning process data, Finally, an accurate teaching early warning is provided to teachers based on a four-level early warning evaluation rule. In experiments, the real datasets are used and the results reveal that the proposed approach can accurately diagnose students' cognitive status and effectively predict teachers' teaching quality. This approach can provide an accurate teaching early warning service for teachers.
Unsupervised domain adaptation (UDA) attracts extra attention in medical image processing because no additional labels are required when adapting to different distributions. In this work, we propose a novel unsupervis...
详细信息
Unsupervised domain adaptation (UDA) attracts extra attention in medical image processing because no additional labels are required when adapting to different distributions. In this work, we propose a novel unsupervised domain adaptation framework named as Domain Expansion and PseudoLabeling (DEPL). We extend the domain of a labeled source domain data to four different distributed domains and use adversarial learning to align the image appearance level and feature level from the four different domains to the unlabeled target domain. In addition, we propose a selective pseudolabeling mechanism, namely using strong confidence pseudolabeling to boost model performance. We evaluate our model for the MR to CT adaptation segmentation task on the public dataset MMWHS. Compared to seven other state-of-the-art segmentation methods, our DEPL achieves the best Dice similarity coefficient by 82.4%, which is at least 3.9% higher than the other UDA segmentation methods.
Within the kinetic energy driven superconducting mechanism, we have studied the temperature dependence of commensurate magnetic resonance in cuprate superconductors. It is shown that the commensurate magnetic resonanc...
详细信息
Aimed at the segmentation problem of ultrawideband synthetic aperture radar (UWB SAR) image, a novel algorithm based on polynomial analysis of statistical distribution is proposed in this letter. Firstly, we estimate ...
详细信息
Large language Models (LLMs) have recently demonstrated remarkable performance in general tasks across various fields. However, their effectiveness within specific domains such as drug development remains challenges. ...
详细信息
In this paper, we propose a novel convolutional neural network based on boundary-guided and region-aware (BGRA-Net) for breast tumor segmentation in ultrasound images. In particular, in the encoding stage, we propose ...
详细信息
ISBN:
(纸本)9781665429825
In this paper, we propose a novel convolutional neural network based on boundary-guided and region-aware (BGRA-Net) for breast tumor segmentation in ultrasound images. In particular, in the encoding stage, we propose a boundary-guided module (BGM) to guide the learning of boundary features in the decoding stage by explicitly strengthening the extracted boundary information. Meanwhile, in the decoding stage, we propose a region-aware module (RAM) to integrate different levels of detailed and semantic features to improve the comprehensive representation of tumor regional features. Besides, a scale-adaptive module (SAM) is further proposed to capture the characteristics of tumors with different sizes between the encoding and decoding stages. To evaluate the effectiveness of our BGRA-Net, we conduct extensive experiments on the UDIAT dataset and compare it with eight state-of-the-art methods. The experimental results show that our BGRA-Net outperforms the state-of-the-art methods and can achieve accurate segmentation of breast tumors with ambiguous boundaries.
Existing machine learning algorithms face the problems of label-missing and high dimensionality. Feature selection is an effective dimensionality reduction method that can improve the efficiency and accuracy of subseq...
详细信息
Distance function is a main metrics of measuring the affinity between two data points in machine learning. Extant distance functions often provide unreachable distance values in real applications. This can lead to inc...
详细信息
Unmanned aerial vehicles (UAV) can be applied in many Internet of Things (IoT) systems, e.g., smart farms, as a data collection platform. However, the UAV-IoT wireless channels may be occasionally blocked by trees or ...
详细信息
暂无评论