检索结果-内蒙古大学图书馆

IAENG International Journal of computer science 2024年第8期51卷 1020-1026页

作者： Miao, Liheng Tian, Ying School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China

The underwater environment is complex and diverse, making it challenging to locate aquatic organisms accurately. The precise identification of underwater animals is crucial for ecological research and fisheries management. Addressing the issue of inaccurate localization of small underwater targets, this study introduces a novel model, YOLOv8-2PCC, based on the YOLOv8 algorithm with improvements. First, to improve the efficiency of the YOLOv8 network, the C2F module in the original YOLOv8 network model was replaced with convolution to reduce the computational load of the model. Secondly, the up-sampling operator CARAFE is employed, which excels in capturing features at various scales. Finally, a small target detection layer has been incorporated to extract additional shallow features, effectively enhancing the model's ability to detect small targets. Utilizing the URPC dataset for training and testing, the results indicate that our proposed algorithm achieves a mean Average Precision (mAP) of 85.9%. Compared to YOLOv8n, there is a 4.4% improvement, effectively enhancing the accuracy of underwater organism detection in complex underwater environments. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Aquatic organisms

来源：评论

学校读者我要写书评

暂无评论

KD-Crowd:a knowledge distillation framework for learning from crowds

引用

Frontiers of computer science 2025年第1期19卷 119-130页

作者： Shaoyuan LI Yuxiang ZHENG Ye SHI Shengjun HUANG Songcan CHEN College of Computer Science and Technology Nanjing University of Aeronautics and AstronauticsMIIT Key Laboratory of Pattern Analysis and Machine IntelligenceNanjing 211106China

Recently, crowdsourcing has established itself as an efficient labeling solution by distributing tasks to crowd workers. As the workers can make mistakes with diverse expertise, one core learning task is to estimate each worker’s expertise, and aggregate over them to infer the latent true labels. In this paper, we show that as one of the major research directions, the noise transition matrix based worker expertise modeling methods commonly overfit the annotation noise, either due to the oversimplified noise assumption or inaccurate estimation. To solve this problem, we propose a knowledge distillation framework (KD-Crowd) by combining the complementary strength of noise-model-free robust learning techniques and transition matrix based worker expertise modeling. The framework consists of two stages: in Stage 1, a noise-model-free robust student model is trained by treating the prediction of a transition matrix based crowdsourcing teacher model as noisy labels, aiming at correcting the teacher’s mistakes and obtaining better true label predictions;in Stage 2, we switch their roles, retraining a better crowdsourcing model using the crowds’ annotations supervised by the refined true label predictions given by Stage 1. Additionally, we propose one f-mutual information gain (MIG^(f)) based knowledge distillation loss, which finds the maximum information intersection between the student’s and teacher’s prediction. We show in experiments that MIG^(f) achieves obvious improvements compared to the regular KL divergence knowledge distillation loss, which tends to force the student to memorize all information of the teacher’s prediction, including its errors. We conduct extensive experiments showing that, as a universal framework, KD-Crowd substantially improves previous crowdsourcing methods on true label prediction and worker expertise estimation.

关键词： crowdsourcing label noise worker expertise knowledge distillation robust learning

来源：评论

学校读者我要写书评

暂无评论

Robust domain adaptation with noisy and shifted label distribution

引用

Frontiers of computer science 2025年第3期19卷 25-36页

作者： Shao-Yuan LI Shi-Ji ZHAO Zheng-Tao CAO Sheng-Jun HUANG Songcan CHEN MIIT Key Laboratory of Pattern Analysis and Machine Intelligence College of Computer Science and TechnologyNanjing University of Aeronautics and AstronauticsNanjing 211106China

Unsupervised Domain Adaptation(UDA)intends to achieve excellent results by transferring knowledge from labeled source domains to unlabeled target domains in which the data or label distribution *** UDA methods have acquired great success when labels in the source domain are ***,even the acquisition of scare clean labels in the source domain needs plenty of costs as *** the presence of label noise in the source domain,the traditional UDA methods will be seriously degraded as they do not deal with the label *** this paper,we propose an approach named Robust Self-training with Label Refinement(RSLR)to address the above *** adopts the self-training framework by maintaining a Labeling Network(LNet)on the source domain,which is used to provide confident pseudo-labels to target samples,and a Target-specific Network(TNet)trained by using the pseudo-labeled *** combat the effect of label noise,LNet progressively distinguishes and refines the mislabeled source *** combination with class rebalancing to combat the label distribution shift issue,RSLR achieves effective performance on extensive benchmark datasets.

关键词： unsupervised domain adaptation label noise label distribution shift self-training class rebalancing

来源：评论

学校读者我要写书评

暂无评论

An Encoding-Decoding Framework Based on CNN for circ RNA-RBP Binding Sites Prediction

引用

Chinese Journal of Electronics 2024年第1期33卷 256-263页

作者： Yajing GUO Xiujuan LEI Yi PAN School of Computer Science Shaanxi Normal University Faculty of Computer Science and Control Engineering Shenzhen Institute of Advanced TechnologyChinese Academy of Sciences Department of Computer Science Georgia State University

Predicting RNA binding protein(RBP) binding sites on circular RNAs(circ RNAs) is a fundamental step to understand their interaction mechanism. Numerous computational methods are developed to solve this problem, but they cannot fully learn the features. Therefore, we propose circ-CNNED, a convolutional neural network(CNN)-based encoding and decoding framework. We first adopt two encoding methods to obtain two original matrices. We preprocess them using CNN before fusion. To capture the feature dependencies, we utilize temporal convolutional network(TCN) and CNN to construct encoding and decoding blocks, respectively. Then we introduce global expectation pooling to learn latent information and enhance the robustness of circ-CNNED. We perform circ-CNNED across 37 datasets to evaluate its effect. The comparison and ablation experiments demonstrate that our method is superior. In addition, motif enrichment analysis on four datasets helps us to explore the reason for performance improvement of circ-CNNED.

关键词： Circular RNAs (circRNAs) RNA binding proteins Convolutional neural network Temporal convolutional network Encoder-decoder network

来源：评论

学校读者我要写书评

暂无评论

State space representation and phase analysis of gradient descent optimizers

引用

science China(Information sciences) 2023年第4期66卷 140-154页

作者： Biyuan YAO Guiqing LI Wei WU School of Computer Science and Engineering South China University of Technology School of Computer Wuhan University

Deep learning has achieved good results in the field of image recognition due to the key role of the optimizer in a deep learning network. In this work, the optimizers of dynamical system models are established,and the influence of parameter adjustments on the dynamic performance of the system is proposed. This is a useful supplement to the theoretical control models of optimizers. First, the system control model is derived based on the iterative formula of the optimizer, the optimizer model is expressed by differential equations, and the control equation of the optimizer is established. Second, based on the system control model of the optimizer, the phase trajectory process of the optimizer model and the influence of different hyperparameters on the system performance of the learning model are analyzed. Finally, controllers with different optimizers and different hyperparameters are used to classify the MNIST and CIFAR-10 datasets to verify the effects of different optimizers on the model learning performance and compare them with related methods. Experimental results show that selecting appropriate optimizers can accelerate the convergence speed of the model and improve the accuracy of model recognition. Furthermore, the convergence speed and performance of the stochastic gradient descent(SGD) optimizer are better than those of the stochastic gradient descent-momentum(SGD-M) and Nesterov accelerated gradient(NAG) optimizers.

关键词： optimizer control model phase trajectory parameter adjustment classification dynamic performance

来源：评论

学校读者我要写书评

暂无评论

Regulation of metal doping and grain size for Cr2O3 to accomplish hyperspectral camouflage properties with high near-infrared reflectance, large red-edge slope, and low visible reflectance

引用

Ceramics International 2025年第17期51卷 23319-23329页

作者： Lin, Litao Guo, Tengchao Chen, Yu Lu, Xiaochi Quan, Bin Huang, Xiaogu School of Chemistry and Materials Science Nanjing University of Information Science and Technology Nanjing210044 China Nanjing University of Posts and Telecommunications Nanjing210023 China

The advancement of hyperspectral detection technology has challenged the development of hyperspectral camouflage materials. Despite the high spectral similarity of existing bio-inspired hyperspectral materials, their poor weather resistance limits their application in harsh environments. To address this, it is essential to enhance the weather resistance of hyperspectral materials while maintaining high spectral similarity. Therefore, inorganic hyperspectral materials are employed to improve durability. However, these inorganic materials generally exhibit lower spectral similarity and insufficient near-infrared reflectance, which restricts their ability to match the spectral characteristics of diverse plant backgrounds. To enhance spectral and color similarity with green plants, this study investigates Cr2-xTixO3, which exhibits low visible reflectance and high near-infrared reflectance. The impact of annealing temperature on reflectance (400–1300 nm) was assessed using SEM, XRD, XPS, and UV–Vis–NIR spectroscopy. The influence of free carrier concentration and particle size on reflectance was analyzed. Hyperspectral coatings based on Cr2-xTixO3 were prepared, demonstrating tunable hyperspectral reflectance. Results indicate that Cr2-xTixO3 annealed at 1100 °C has the highest near-infrared reflectance, largest red-edge slope, and tunable range, with low visible reflectance. Ultimately, the optimized spectral similarity of these coatings can reach 97 %, coupled with superior environmental durability. © 2025 Elsevier Ltd and Techna Group S.r.l.

关键词： Camouflage

来源：评论

学校读者我要写书评

暂无评论

Cocrystal@protein-anchoring nanococktail for combinatorially treating multidrug-resistant cancer

引用

Acta Pharmaceutica Sinica B 2024年第10期14卷 4509-4525页

作者： Jiahui Zou Xuyang Xing Chao Teng Qingling Zhao Wei He Xuri Wu Yuanzheng Xia School of Pharmacy China Pharmaceutical UniversityNanjing 211198China School of Computer Science and Engineering Nanjing University of Science and TechnologyNanjing 210094China School of Life Science and Technology China Pharmaceutical UniversityNanjing 211198China State Key Laboratory of Natural Medicines and Jiangsu Key Laboratory of Bioactive Natural Product Research School of Traditional Chinese PharmacyChina Pharmaceutical UniversityNanjing 211198China

Multidrug resistance(MDR),the major mechanism by which various cancers develop specific resistance to therapeutic agents,has set up enormous obstacles to many forms of tumor *** cocktail therapy administration,based on the combination of multiple drugs for anti-MDR chemotherapy,often suffers from inconsistent in vivo pharmacokinetic behaviors that cannot act synchronously on the lesions,leading to limited pharmacodynamic *** the emergence of nanomedicines,which has improved chemotherapeutic drugs’bioavailability and therapeutic effect on clinical application,these monotherapy-based nano-formulations still show poor progression in overcoming ***,a“one stone and three birds”nanococktail integrated by a cocrystal@protein-anchoring strategy was purposed for triple-payload delivery,which paclitaxel-disulfiram cocrystal-like nanorods(NRs)were anchored with the basic protein drug Cytochrome c(Cyt C),followed by hyaluronic-acid *** particular,NRs were utilized as carrier-like particles to synchronously deliver biomacromolecule Cyt C into tumor cells and then promote cell *** note,on A549/Taxol drug-resistant tumor-bearing mice,the system with extraordinarily high encapsulation efficiency demonstrated prolonged in vivo circulation and increased tumor-targeting accumulation,significantly reversing tumor drug resistance and improving therapeutic *** mechanistic study indicated that the system induced the apoptosis of Taxol-resistant tumor cells through the signal axis P-glycoprotein/Cyt C/caspase ***,this nanococktail strategy offers a promising approach to improve the sensitivity of tumor cells to chemotherapeutic drugs and strengthen intractable drug-resistant oncotherapy.

关键词： Cocktail Multidrug-resistant Co-delivery combined therapy Cytochrome C Nanocrystals

来源：评论

学校读者我要写书评

暂无评论

Infrared Road Object Detection Based on Improved YOLOv8

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer science 2024年第3期51卷 252-259页

作者： Luo, Zilong Tian, Ying School of Computer Science and Software Engineering University of Science and Technology Liaoning Liaoning 114051 China School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China

In recent years, infrared target detection has played a crucial role in intelligent transportation and assisted driving. Addressing the current issues of low detection accuracy, poor robustness, and missed detections in infrared image detection, we propose an improved infrared road traffic detection algorithm, YOLOv8-EGP, based on YOLOv8s. Firstly, we replace the original C2f module with the SCConv convolution module to extract feature information of different sizes, thereby enhancing the perception of local features in infrared images and connecting spatial relationships between them. Then, in the head part, we use the dyhead detection head and combine three different dimensions with multi-attention, improving the expression ability of the detection head for infrared targets without increasing computational complexity. Finally, we add a small target detection layer (min) to reduce missed detections of small targets in infrared images and improve the final detection accuracy. The conducted ablation experiments show that on the FILIR public dataset, compared to YOLOv8s, the YOLOv8-EGP algorithm increases mAP50 by 6.1%, and precision and recall also increase by 5.8% and 1.6%, respectively, indicating that the improved model can better adapt to infrared target detection, validating the effectiveness of this method. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Research on Social Recommendation Algorithm Based on PSO_KFCM Clustering and CBAM Attention Mechanism of Graph Neural Networks

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer science 2024年第8期51卷 936-948页

作者： Teng, Yue Yang, Kai School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan114051 China

In today's society, people increasingly need information acquisition due to the rapid development of science and technology and the consequent increase in available data. However, finding the information users need from this vast data has become challenging. To tackle this problem, recommending preferred information to users is becoming increasingly important. However, accurately recommending information by analyzing existing models such as GraphRec is still a challenging problem. A method called PSO_KFCM is proposed in this paper to solve this problem better. The technique combines Particle Swarm Optimization (PSO) with hybrid optimization and the kernel fuzzy C-means clustering technique to cluster similar recommendation data into one class. This way, the complexity and randomness of the recommendation data are reduced. It improves the speed and accuracy of the model prediction, which lays a solid foundation for the subsequent recommendation. Various factors will impact the recommendation process, and channel and spatial characteristics are essential. CBAM attention is added to the original attention mechanism to fully utilize these features in the recommendation data to enhance its performance. Furthermore, this paper proposes a social recommendation prediction method that combines CBAM attention and PSO_KFCM clustering and introduces a new social model called TTYGNN. The TTYGNN model optimizes the recommendation effect while maintaining the original advantages, enabling users to obtain the required information more quickly and accurately. To verify the effectiveness and practicality of the proposed model, extensive experimental comparisons were conducted on two widely used datasets. The results show that the TTYGNN model outperforms similar methods in all indicators, proving its superiority in information recommendation. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Fuzzy clustering

来源：评论

学校读者我要写书评

暂无评论

Multi-Task Chinese Speech Recognition Method Based on the Squeezeformer Model

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer science 2025年第1期52卷 23-31页

作者： Guo, Ying Wang, Li School of Computer Science and software Engineering University of Science and Technology Liaoning Anshan114051 China College of Computer Science and Technology Liaoning Anshan114051 China

End-to-end training has emerged as a prominent trend in speech recognition, with Conformer models effectively integrating Transformer and CNN architectures. However, their complexity and high computational cost pose deployment challenges. To address these issues, we propose a multi-task Chinese speech recognition method based on the Squeezeformer model. We replace the FMCF structure in Conformer with an MF/CF structure, leveraging the convolutional module as a local Multi- Head Attention (MHA) module to enhance efficiency. Multilevel down-sampling and up-sampling using a time-series U-Net further reduce computational costs. By eliminating redundant LayerNorm layers and employing depthwise separable convolutions, we streamline the model, reduce parameters, and lower deployment costs. An Adaptor Layer is integrated into the MHSA module to mitigate the vanishing gradient problem, and a ScaleVar Layer is added to enhance flexibility. Additionally, the RealFormer module is introduced on the decoding side to improve context understanding. Combining Connectionist Temporal Classification (CTC) with attention-based encoding and decoding models for multi-task learning improves performance and accuracy. Experimental results show that the proposed method reduces the parameters on AISHELL-1 dataset by 16% and reduces the character error rate to 5.50%. At the same time, it also shows good performance on AISHELL-2 dataset. © (2025), (International Association of Engineers). All rights reserved.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：