检索结果-内蒙古大学图书馆

SSRN 2023年

作者： Miao, Qing Wu, Xiaohe Xu, Chao Ji, Yanli Zuo, Wangmeng Guo, Yiwen Meng, Zhaopeng College of Intelligence and Computing Tianjin University Tianjin300072 China School of Computer Science and Technology Harbin Institute of Technology Harbin150006 China Center for Future Media School of Computer Science and Engineering UESTC Chengdu611731 China China

Learning with noisy labels (LNL) has been extensively studied, with existing approaches typically following a framework that alternates between clean sample selection and semi-supervised learning (SSL). However, this approach has a limitation: the clean set selected by the Deep Neural Network (DNN) classifier, trained through self-training, inevitably contains noisy samples. This mixture of clean and noisy samples leads to misguidance in DNN training during SSL, resulting in impaired generalization performance due to confirmation bias caused by error accumulation in sample selection. To address this issue, we propose a method called Collaborative Sample Selection (CSS), which leverages the large-scale pre-trained model CLIP. CSS aims to remove the mixed noisy samples from the identified clean set. We achieve this by training a 2-Dimensional Gaussian Mixture Model (2D-GMM) that combines the probabilities from CLIP with the predictions from the DNN classifier. To further enhance the adaptation of CLIP to LNL, we introduce a co-training mechanism with a contrastive loss in semi-supervised learning. This allows us to jointly train the prompt of CLIP and the DNN classifier, resulting in improved feature representation, boosted classification performance of DNNs, and reciprocal benefits to our Collaborative Sample Selection. By incorporating auxiliary information from CLIP and utilizing prompt fine-tuning, we effectively eliminate noisy samples from the clean set and mitigate confirmation bias during training. Experimental results on multiple benchmark datasets demonstrate the effectiveness of our proposed method in comparison with the state-of-the-art approaches. © 2023, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A lightweight method for Android malware classification based on teacher assistant distillation

A lightweight method for Android malware classification base...

引用

International Conference on Mobile Ad-hoc and Sensor Networks, MSN

作者： Junwei Tang Qiaosen Pi Ruhan He Jin Huang Tao Peng Xinrong Hu Wenlong Tian Hubei Provincial Engineering Research Center for Intelligent Textile and Fashion Wuhan Textile University School of Computer Science and Artificial Intelligence Wuhan Textile University School of Computer Science University of South China School of Physical and Mathematical Sciences Nanyang Technological University Singapore

ISBN: (数字)9798350358261

ISBN: (纸本)9798350358278

In recent years, the growing concern over mobile security and the associated risks posed by mobile malware have prompted an increased focus on utilizing deep learning models for analyzing Android application security. However, the expansion of deep learning model sizes results in an exponential growth of model parameters, demanding significant computing resources for execution. To address this challenge, we propose a lightweight Android malware detection method based on teacher-assistant-student knowledge distillation. Our method enables predicting on local clients, eliminating the need for cloud-base service interactions, and protecting user privacy. We visualize the binary file of the target Android application as an RGB three-channel color image, using ResNeSt50 as the teacher model, and compress it based on knowledge distillation. An assistant model is incorporated to address the issue of insufficient distillation resulting from the significant gap between the teacher and student models. Additionally, we integrate a split-attention mechanism to enhance the ability of the professor model to acquire deep features of malware images. We conduct experiments on Drebin and CICMalDroid 2020 datasets and the results show that the proposed method can ensure that the detection results of student model are more similar to those of the teacher model while reducing model complexity. Our method reduces the number of model parameters by 95% compare to the teacher model while maintaining accuracy. And the accuracy is improved by 0.63% compare to the traditional distillation method.

关键词： Deep learning Visualization Privacy Accuracy Computational modeling Transforms Malware

来源：评论

学校读者我要写书评

暂无评论

Pedestrain Detection in Infrared Images with Improved YOLOv2 Network

Pedestrain Detection in Infrared Images with Improved YOLOv2...

引用

2020 IEEE International Conference on Information Technology, Big Data and Artificial Intelligence, ICIBA 2020

作者： Du, Chenshi Song, Peng Ma, Xin Center for Robotics School of Control Science and Engineering Shandong University Jinan China

ISBN: (纸本)9781728152240

With the progress of unmanned driving technology and the attention of security problems, pedestrian detection is getting more and more researchers' participation. Since 2012, the application of deep convolutional network has greatly improved the accuracy and detection speed of pedestrian detection. However, it is difficult to detect pedestrians at night, the detection accuracy of small objects and obstructed pedestrians is poor, which are problems that researchers are solving. This paper presents infrared image detection of pedestrain based on improved YOLOv2 detection network. It improves the detection accuracy of small volume and blocked pedestrian in the night. © 2020 IEEE.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Facial Action Unit Detection by Adaptively Constraining Self-Attention and Causally Deconfounding Sample

arXiv

引用

arXiv 2024年

作者： Shao, Zhiwen Zhu, Hancheng Zhou, Yong Xiang, Xiang Liu, Bing Yao, Rui Ma, Lizhuang School of Computer Science and Technology China University of Mining and Technology Xuzhou221116 China Ministry of Education Wuhan430074 China Mine Digitization Engineering Research Center of the Ministry of Education Xuzhou221116 China Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai200240 China

Facial action unit (AU) detection remains a challenging task, due to the subtlety, dynamics, and diversity of AUs. Recently, the prevailing techniques of self-attention and causal inference have been introduced to AU detection. However, most existing methods directly learn self-attention guided by AU detection, or employ common patterns for all AUs during causal intervention. The former often captures irrelevant information in a global range, and the latter ignores the specific causal characteristic of each AU. In this paper, we propose a novel AU detection framework called AC2D by adaptively constraining self-attention weight distribution and causally deconfounding the sample confounder. Specifically, we explore the mechanism of self-attention weight distribution, in which the self-attention weight distribution of each AU is regarded as spatial distribution and is adaptively learned under the constraint of location-predefined attention and the guidance of AU detection. Moreover, we propose a causal intervention module for each AU, in which the bias caused by training samples and the interference from irrelevant AUs are both suppressed. Extensive experiments show that our method achieves competitive performance compared to state-of-the-art AU detection approaches on challenging benchmarks, including BP4D, DISFA, GFT, and BP4D+ in constrained scenarios and Aff-Wild2 in unconstrained scenarios. The code is available at https://***/ZhiwenShao/AC2D. Copyright © 2024, The Authors. All rights reserved.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Cross-Lingual Summarization: English - Bahasa Indonesia 6

Cross-Lingual Summarization: English - Bahasa Indonesia

引用

6th International Workshop on Big Data and Information Security, IWBIS 2021

作者： Abka, Achmad F. Pratama, Mahardhika Jatmiko, Wisnu National Research and Innovation Agency Research Center for Informatics Indonesia Nanyang Technological University School of Computer Science and Engineering Singapore Singapore Universitas Indonesia Faculty of Computer Science Indonesia Indonesia

ISBN: (数字)9781665424516

ISBN: (纸本)9781665424516

Progress of abstractive summarization has been accelerated since the introduction of sequence-to-sequence neural networks. Summarization is no longer limited to selecting words or sentences that exist in the source document as in the extractive approach, but can generate completely new words or sentences that have never appeared in the source document. Big push came from machine translation research with the introduction of attention mechanisms. Attention mechanism is the key to the information bottleneck problem in encoder-decoder model. Cross-Lingual Summarization (CLS) is the task of generating a summary in target language from source document in different language. Traditional methods split this task into two steps: summarization and translation. This paper describes a study on CLS without explicitly using translator, thereby reducing one step as in existing method. We incorporate multilingual embeddings in sequence-to-sequence neural networks with attention mechanisms to handle this task. Multilingual embeddings are used to represent words as if the source language and the target language are the same language. Experiments show comparable performance between monolingual summarization and cross-lingual summarization in Amazon Fine Food review data indicated by ROUGE scores which are only 1-2 points apart. © 2021 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Iterative Shaping of Multi-Particle Aggregates based on Action Trees and VLM

arXiv

引用

arXiv 2025年

作者： Lee, Hoi-Yin Zhou, Peng Duan, Anqing Yang, Chenguang Navarro-Alarcon, David Department of Mechanical Engineering The Hong Kong Polytechnic University KLN Hong Kong School of Advanced Engineering The Great Bay University Dongguan China Department of Robotics Mohamed Bin Zayed University of Artificial Intelligence United Arab Emirates Department of Computer Science University of Liverpool Liverpool United Kingdom

In this paper, we address the problem of manipulating multi-particle aggregates using a bimanual robotic system. Our approach enables the autonomous transport of dispersed particles through a series of shaping and pushing actions using robotically-controlled tools. Achieving this advanced manipulation capability presents two key challenges: high-level task planning and trajectory execution. For task planning, we leverage Vision Language Models (VLMs) to enable primitive actions such as tool affordance grasping and non-prehensile particle pushing. For trajectory execution, we represent the evolving particle aggregate's contour using truncated Fourier series, providing efficient parametrization of its closed shape. We adaptively compute trajectory waypoints based on group cohesion and the geometric centroid of the aggregate, accounting for its spatial distribution and collective motion. Through realworld experiments, we demonstrate the effectiveness of our methodology in actively shaping and manipulating multi-particle aggregates while maintaining high system cohesion. © 2025, CC BY-NC-SA.

关键词： Fourier series

来源：评论

学校读者我要写书评

暂无评论

A Multi-Resolution Mutual Learning Network for Multi-Label ECG Classification

A Multi-Resolution Mutual Learning Network for Multi-Label E...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Wei Huang Ning Wang Panpan Feng Haiyan Wang Zongmin Wang Bing Zhou School of Computer and Artificial Intelligence Zhengzhou University Zhengzhou China Cooperative Innovation Center of Internet Healthcare Zhengzhou University Zhengzhou China School of Communication and Information Engineering Nanjing University of Posts and Telecommunications Nanjing China School of Computer Science Zhengzhou University of Aeronautics Zhengzhou China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Electrocardiograms (ECG), essential for diagnosing cardiovascular diseases. In recent years, the application of deep learning techniques has significantly improved the performance of ECG signal classification. Multi-resolution feature analysis, which captures and processes information at different time scales, can extract subtle changes and overall trends in ECG signals, showing unique advantages. However, common multi-resolution analysis methods based on simple feature addition or concatenation may lead to the neglect of low-resolution features, affecting model performance. To address this issue, this paper proposes the Multi-Resolution Mutual Learning Network (MRMNet). MRM-Net includes a dual-resolution attention architecture and a feature complementary mechanism. The dual-resolution attention architecture processes high-resolution and low-resolution features in parallel. Through the attention mechanism, the high-resolution and low-resolution branches can focus on subtle waveform changes and overall rhythm patterns, enhancing the ability to capture critical features in ECG signals. Meanwhile, the feature complementary mechanism introduces mutual feature learning after each layer of the feature extractor. This allows features at different resolutions to reinforce each other, thereby reducing information loss and improving model performance and robustness. Experiments on the PTB-XL and CPSC2018 datasets demonstrate that MRM-Net significantly outperforms existing methods in multi-label ECG classification performance. The code for our framework will be publicly available at https://***/wxhdf/MRM.

关键词： Representation learning Attention mechanisms Pattern classification Electrocardiography Feature extraction Rhythm Market research Robustness Data mining Signal resolution

来源：评论

学校读者我要写书评

暂无评论

A Machine Learning-Based Forecasting Model with Deep Learning Capabilities for Unbalanced Time Series Data 2

A Machine Learning-Based Forecasting Model with Deep Learnin...

引用

2nd IEEE Global Conference on Communications and Information Technologies, GCCIT 2024

作者： Amudha, A. Suri, Sahil Sindhu, M. Goyal, Shubhi Kukreja, Mansi Pathak, Kishor R. Karpagam Academy of Higher Education Department of Electrical and Electronics Engineering Coimbatore641021 India Chitkara University Centre of Research Impact and Outcome Punjab Rajpura140417 India Prince Shri Venkateshwara Padmavathy Engineering College Department of Computer Science and Engineering Chennai127 India Quantum University Research Center Quantum University Uttarakhand Roorkee247667 India School of Mangement - Pg Management Karnataka Bangalore India Vishwakarma Institute of Technology Department of Information Technology Pune India

ISBN: (纸本)9798350388916

During the last several years, deep learning capabilities have surpassed those of more conventional models on a wide variety of machine learning-related tasks. Neural networks with deep layers have indeed been effectively used to time series forecasting issues, which are crucial to the field of data mining. Their ability to instantly understand the temporal connections contained in time series has made them a viable option. Yet, deciding on a deep neural network architecture and then optimizing its parameters calls for a high level of skill. Thus, it is essential to conduct comprehensive studies of the predictive capacities of all current architectural frameworks. The completion of this work is hampered by two primary factors: (1) a complete examination of the most recent research on depth learning's application to time series forecasting, and (2) an empirical inquiry analysing the effectiveness of the most commonly used architectures. Seven distinct deep learning models are evaluated and compared for their efficiency and accuracy. We present scores and distributions based on our analysis of the suggested models trained with various architectures and hyperparameter settings. The 50,000 time series in the databases are used to solve 12 unique forecasting issues. Using this data to train approximately 38,000 models, we present the largest deep learning research to date for forecasting time series. While LSTMs were shown to be the most significantly predictive, both LSTMs & CNNs showed promise in this research. Comparable performance is attained by CNNs, but with less variance in results as well as a more based on the best available overall. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Exploring CLIP’s Dense Knowledge for Weakly Supervised Semantic Segmentation

arXiv

引用

arXiv 2025年

作者： Yang, Zhiwei Meng, Yucong Fu, Kexue Tang, Feilong Wang, Shuo Song, Zhijian Academy for Engineering and Technology Fudan University Shanghai200433 China Shanghai Key Laboratory of Medical Image Computing and Computer Assisted Intervention China Digital Medical Research Center School of Basic Medical Sciences Fudan University China Shandong Computer Science Center National Supercomputer Center in Jinan China

Weakly Supervised Semantic Segmentation (WSSS) with image-level labels aims to achieve pixel-level predictions using Class Activation Maps (CAMs). Recently, Contrastive Language-Image Pre-training (CLIP) has been introduced in WSSS. However, recent methods primarily focus on image-text alignment for CAM generation, while CLIP’s potential in patch-text alignment remains unexplored. In this work, we propose ExCEL to explore CLIP’s dense knowledge via a novel patch-text alignment paradigm for WSSS. Specifically, we propose Text Semantic Enrichment (TSE) and Visual Calibration (VC) modules to improve the dense alignment across both text and vision modalities. To make text embeddings semantically informative, our TSE module applies Large Language Models (LLMs) to build a dataset-wide knowledge base and enriches the text representations with an implicit attribute-hunting process. To mine fine-grained knowledge from visual features, our VC module first proposes Static Visual Calibration (SVC) to propagate fine-grained knowledge in a non-parametric manner. Then Learnable Visual Calibration (LVC) is further proposed to dynamically shift the frozen features towards distributions with diverse semantics. With these enhancements, ExCEL not only retains CLIP’s training-free advantages but also significantly outperforms other state-of-the-art methods with much less training cost on PASCAL VOC and MS COCO. Code is available at https://***/zwyang6/ExCEL. Copyright © 2025, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

An Optimization-Aware Pre-Routing Timing Prediction Framework Based on Multi-Modal Learning

引用

IEEE Transactions on computer-Aided Design of Integrated Circuits and Systems 2025年

作者： Cao, Peng Qin, Yusen He, Guoqing Ding, Wenjie Cheng, Xu Zhang, Zhanhua Ye, Yuyang Southeast University School of Integrated Circuits Nanjing210096 China National Center of Technology Innovation for Eda Nanjing210031 China The Chinese University of Hong Kong Department of Computer Science and Engineering Hong Kong Hong Kong

Accurate and efficient pre-routing timing estimation is particularly crucial during placement to alleviate time-consuming design iterations. Machine learning (ML) based methods have been introduced recently to predict the post-routing timing results at placement stage, but most of them neglect the impact of timing optimization during physical design, suffering from accuracy loss due to inconsistent circuit netlist. In this work, an optimization-aware pre-routing timing prediction framework based on multi-modal learning is proposed to calibrate the timing changes between placement and routing stages, where the local netlist and layout information are extracted by graph neural network (GNN) and convolutional neural network (CNN) respectively while the global information along the path is further extracted by Transformer network. Based on the predicted post-routing timing results by the proposed framework, timing optimization guidance is generated to enhance traditional design flow with better physical implementation quality. Experimental results demonstrate that for the OpenCores benchmark circuits under TSMC 22nm process, the proposed framework achieves significant correlation and accuracy improvement with an average of 0.9219 in terms of R score and 2.12% of mean absolute percentage error (MAPE) as well as an average runtime acceleration of 645× compared with traditional design flow on testing designs. With the timing optimization guidance, significant worst negative slack (WNS) and total negative slack (TNS) improvement are achieved compared with traditional flow after placement and routing respectively without noticeable area, power, wire length, and the number of design rule check (DRC) violations increase. © 2025 IEEE.

关键词： Convolutional neural networks

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：