检索结果-内蒙古大学图书馆

2025 IEEE international conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Shi, Jianxin Chen, Jinhao Chen, Xiaolong Ma, Jun Wo, Tianyu School of Computer Science and Engineering Beihang University Beijing China Robotics and Autonomous Systems Thrust The Hong Kong University of Science and Technology Guangzhou China

ISBN: (纸本)9798350368741

Accurately predicting the future motions of traffic agents is essential for autonomous systems. Despite the significant success of existing motion forecasting methods based on supervised learning, they still exhibit two main limitations. First, when annotated data for a scene is limited, these methods often fail to achieve the expected accuracy. Second, they typically rely on complex architectures and extensive prior knowledge to improve performance. To overcome these challenges, we propose MF-BERT, a novel framework that adapts the concept of BERT to motion forecasting, inspired by advancements in the self-supervised pre-training paradigm. During pre-training, we design a siamese sequence modeling task with an asymmetric mask strategy to capture complex behavior patterns of agents. During fine-tuning, the pre-trained representation module initializes the feature encoder of the motion forecasting model, and a multimodal trajectory decoder generates all possible predictions. Experimental results demonstrate the superiority of MF-BERT over state-of-the-art methods. © 2025 IEEE.

关键词： Autonomous Motion Forecasting Self-supervised learning Time Series Analysis

来源：评论

学校读者我要写书评

暂无评论

TSR-Jack: An In-Browser Crypto-Jacking Detection Method Based on Time Series Representation learning 26th

TSR-Jack: An In-Browser Crypto-Jacking Detection Method Base...

引用

26th international conference on Information and Communications Security, ICICS 2024

作者： Cui, Bo Liu, Shuai Engineering Research Center of Ecological Big Data Ministry of Education Beijing China College of Computer Science Inner Mongolia University Hohhot China

ISBN: (纸本)9789819788002

As the economic value of cryptocurrencies continues to ascend, an increasing number of cybercriminals exploit malicious browser scripts to commandeer the system and network resources of victims for unauthorized cryptocurrency mining. This form of browser-based mining, termed crypto-jacking, imparts substantial impact and harm to society, enterprises, and users. However, existing research methods lead to poor transferability because they require model training using labeling data from detection devices. To solve this problem, we introduce a crypto-jacking detection method based on time series representation learning, called TSR-jack. Specifically, we first collect hardware runtime system data for a period of time and generate datasets. After slicing the time series datasets, they are used to train a time series representation model, which introduces the loss of temporal information to extract temporal features. The temporal features can capture similarities between different devices and therefore address model transfer difficulties. Finally, the time vectors obtained from representation learning are fed into SVM for classification. Experimental results show that our TSR-jack method outperforms similar dynamic feature detection methods, especially when considering transferability, with a precision improvement of more than 5%. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Plaintext-Free Deep learning for Privacy-Preserving Medical Image Analysis through Frequency Information Embedding

Plaintext-Free Deep Learning for Privacy-Preserving Medical ...

引用

2025 IEEE international conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Sun, Mengyu Yang, Ziyuan Ran, Maosong Wang, Zhiwen Yu, Hui Zhang, Yi School of Cyber Science and Engineering Sichuan University Chengdu China School of Computer Science Sichuan University Chengdu China

ISBN: (纸本)9798350368741

In the fast-evolving field of medical image analysis, deep learning (DL)-based methods have achieved tremendous success. However, these methods require plaintext data for training and inference stages, raising privacy concerns, especially in the sensitive area of medical data. To tackle these concerns, this paper proposes a novel framework that uses surrogate images for analysis, eliminating the need for plaintext images. This approach is called Frequency-domain Exchange Style Fusion (FESF). The framework includes two main components: Image Hidden Module (IHM) and Image Quality Enhancement Module (IQEM). The IHM performs in the frequency domain, blending the features of plaintext medical images into host medical images, and then combines this with IQEM to improve and create surrogate images effectively. During the diagnostic model training process, only surrogate images are used, enabling anonymous analysis without any plaintext data during both training and inference stages. Extensive experiments demonstrate that our framework effectively preserves the privacy of medical images and maintains the diagnostic accuracy of DL models at a relatively high level, proving its effectiveness across various datasets and DL-based models. © 2025 IEEE.

关键词： Deep learning Frequency-Domain Exchange Image Concealment Medical Image Analysis Privacy Preserving

来源：评论

学校读者我要写书评

暂无评论

Quality of AI Service Assurance in 6G Native Artificial Intelligence Networks 3rd

Quality of AI Service Assurance in 6G Native Artificial Inte...

引用

3rd international conference on machine learning, Cloud Computing and Intelligent Mining, MLCCIM 2024

作者： Hua, Meihui Chen, Tianjiao Liu, Guangyi Deng, Juan Li, Na Wang, Xiaoyun China Mobile Research Institute Beijing China ZGC Institute of Ubiquitous-X Innovation and Applications Beijing China China Mobile Communications Group Corporation Beijing China

ISBN: (纸本)9789819624676

6G native artificial intelligence (AI) networks are expected to support diverse vertical industries and offer countless emerging AI services. To satisfy stringent requirements of diversified services, network slicing is developed, which enables service-oriented resource allocation tailored for different AI scenarios. However, the mismatch between the user’s quality of AI service (QoAIS) requirements and the actual capacity of network slices, as well as the hierarchical network architecture, pose great challenges to the guarantee of QoAIS. This paper introduces AI-assisted slicing management framework to maintain a high performance of network operation by supporting diverse AI services and meeting QoAIS. A novel methodology and resource allocation scheme based on federated learning and transfer learning is proposed, which enables efficient life-cycle management of AI services and selection of network slices. Simulation results illustrate that the proposed scheme has superior performance in ensuring QoAIS compared with benchmarks. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

So Far Yet So Near: Time Series data Augmentation with Exploring non-Semantic Boundaries based on Reinforcement learning

So Far Yet So Near: Time Series Data Augmentation with Explo...

引用

2025 IEEE international conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Li, Haoran Zhang, Zhibo Kang, Jiarong Jiang, Xun Gong, Xiaoli Zhang, Jin Sun, Zhe Cichocki, Andrzej College of Computer Science Nankai University Tianjin China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China Faculty of Medicine Faculty of Health Data Science Juntendo University Japan Systems Research Institute Polish Academy of Science Warsaw Poland

ISBN: (纸本)9798350368741

data augmentation effectively expands feature distribution in time series classification, enhancing downstream task performance. However, existing techniques often fail to maintain semantic consistency between augmented and original time series data, causing label noise and thereby degrading downstream task performance. We argue that data augmentation should preserve time series semantic consistency and expand the non-semantic information space. In this paper, we reformulate data augmentation as a semantic path planning problem between original data and augmented data, modeled as a Markov Decision Process (MDP). We propose a reinforcement learning-based algorithm (RL) named FreqSYN, where the action space is defined by a set of learnable Gaussian kernels that perturbs the frequency domain of the original data to generate augmented samples. The confidence coefficients of augmented data in semantically relevant classification tasks are used as a reward to iteratively refine the FreqSYN. Our method is validated across four datasets, achieving state-of-the-art performance, with a 2% improvement in F1 score over the SimPSI method. The code and models are available at https://***/NKU-EmbeddedSystem/FreqSYN. © 2025 IEEE.

关键词： data argumentation frequency domain augmentation Gaussian kernel prediction time series

来源：评论

学校读者我要写书评

暂无评论

CodeQG: Automated Multiple Question Generation for Source Code Comprehension

引用

international JOURNAL OF SOFTWARE engineering AND KNOWLEDGE engineering 2025年第5期35卷 737-765页

作者： Zhang, Xiaowei Chen, Lin Qi, Kaiyuan Zou, Weiqin Pang, Liye Zhang, Lianfa Zhang, Peng Xu, Guanqun Zhang, Dong Inspur Grp Co Ltd Jinan 250101 Peoples R China Jinan Inspur Data Technol Co Ltd Jinan 250101 Peoples R China State Key Lab High End Server Syst Jinan 250101 Peoples R China Nanjing Univ State Key Lab Novel Software Technol Nanjing 210023 Peoples R China Nanjing Univ Aeronaut & Astronaut Nanjing 210016 Peoples R China IEIT Syst Co Ltd Jinan 250101 Peoples R China

During software maintenance and evolution, developers spend more than half of their time on code comprehension activities. In order to understand an unfamiliar code base, they would naturally ask different types of questions related to code snippets and try to find the answers. In this paper, we conduct an initial work to explore the possibility of automatic question generation for program comprehension. We construct a large-scale data set containing pairs of source code and questions that are automatically transformed from inline comments based on dependency analysis and semantic role labeling. We also build a comprehensive taxonomy of question types so as to generate questions concerning different aspects of code snippets, such as purpose, implementation details and so on. Then, we propose a deep learning-based prototype CodeQG to automatically generates multiple types of questions for code snippets. We evaluate CodeQG by using both typical performance metrics and manual evaluation. The results show that (1) we can achieve a value of 42.02 on BLEU4 and 60.81 on ROUGE-L for the generated questions;(2) overall, the questions are very correct in grammatical, semantic and format;(3) the questions are related to the corresponding code snippet and are helpful for developers in source code comprehension activities. Our work gives insights into automatically generating multiple types of questions for code comprehension. We expect this exploration will improve the applicability and generality of machine code comprehension.

关键词： Code question generation code comprehension data sets construction neural networks

来源：评论

学校读者我要写书评

暂无评论

Proceedings - international conference Advancement in data Science, E-learning and Information Systems, ICADEIS 2022

Proceedings - International Conference Advancement in Data S...

引用

4th international conference Advancement in data Science, E-learning and Information Systems, ICADEIS 2022

ISBN: (纸本)9781665463874

The proceedings contain 38 papers. The topics discussed include: developing IT strategic transformation of smart village concept for Indonesian village model;social media analysis on aquaculture supplychain management: a case study on freshwater lobsters;selection of the best color space for image steganography with the least significant bit method;defense-in-depth security strategy in Log4j vulnerability analysis;exploration of ECG-based real-time arrhythmia detection: a systematic literature review;prediction model of mortality with respiratory rate, oxygen saturation and heart rate using logistic regression;application of reverse engineering and simulation in the design of a pontoon drum for coconut shell combustion;optimization of Hijaiyah letter handwriting recognition model based on deep learning;and event-driven architecture to improve performance and scalability in microservices-based systems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Federated learning Framework for Lightweight Model Contrast for Finger Vein Recognition 18th

A Federated Learning Framework for Lightweight Model Contra...

引用

18th Chinese conference on Biometric Recognition, CCBR 2024

作者： Chen, Guang Xie, Tianming Yang, Xu Tian, Feng Kang, Wenxiong School of Automation Science and Engineering South China University of Technology Guangzhou510641 China GRGBanking Equipment Co. Ltd. Guangzhou510663 China

ISBN: (纸本)9789819610679

The training efficacy of finger vein models is influenced by varied device image acquisition characteristics, collector gestures, and contact modes. However, conventional approaches relying on vast data volumes face challenges in the finger vein domain due to stringent security and privacy concerns, limiting data availability. This study proposes a federal learning framework for vein recognition model training, introducing FedFvMCo, a model of contrastive learning among participants. By comparing global and local lightweight model parameters under data convergence limitations, FedFvMCo enhances cross-domain finger vein recognition performance and model generalization. Incorporating an attention mechanism into the lightweight model further optimizes its application in resource-constrained intelligent devices. This framework is validated using four public datasets, demonstrating improved model performance and independence from data convergence constraints. This method can be extended to different modes of biometrics recognition. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

2nd international conference on data Analytics and learning, DAL 2022

2nd International Conference on Data Analytics and Learning,...

引用

2nd international conference on data Analytics and learning, DAL 2022

ISBN: (纸本)9789819963454

The proceedings contain 20 papers. The special focus in this conference is on data Analytics and learning. The topics include: Nitrogen Deficiency and Yield Estimation in Paddy Field;medical Image Compression Using Huffman Coding for Tiff Images;Efficient Wavelet Based Denoising Technique Combined with Features of Cyclespinning and BM3D for Grayscale and Color Images;PBRAMEC: Prioritized Buffer Based Resource Allocation for Mobile Edge Computing Devices;surface Water Quality Analysis Using IoT;children Facial Growth Pattern Analysis Using Deep Convolutional Neural Networks;classification of Forged Logo Images;detection, Classification and Counting of Moving Vehicles from Videos;two-Stage Word Spotting Scheme for Historical Handwritten Devanagari Documents;face Recognition Using Sketch Images;3D Object Detection in Point Cloud Using Key Point Detection Network;IRIS and Face-Based Multimodal Biometrics Systems;a Survey on the Detection of Diseases in Plants Using the Computer Vision-Based Model;an Approach to Conserve Wildlife Habitat by Predicting Forest Fire Using machine learning Technique;A Method to Detect Phishing Websites Using Distinctive URL Characteristics by Employing machine learning Technique;aquaculture Monitoring System: A Prescriptive Model;machine learning-Based Pattern Recognition Models for Image Recognition and Classification;a Review of Silk Farming Automation Using Artificial Intelligence, machine learning, and Cloud-Based Solutions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards Natural Gas Consumption Prediction by Deep learning: Case Study in Several East-Europe Countries 16th

Towards Natural Gas Consumption Prediction by Deep Learning:...

引用

16th international conference on Genetic and Evolutionary Computing, ICGEC 2024

作者： Nguyen, TuanAnh Nguyen, HongGiang Hue University Hue City530000 Viet Nam

ISBN: (纸本)9789819615346

This paper describes a deep learning method such as the long short-term memory (LSTM), bidirectional LSTM (Bi-LSTM), gated recurrent unit (GRU), bidirectional gated recurrent unit (Bi-GRU) to design for forecasting in terms of solving this problem. The data for the study was obtained on the Eurostat website from January 2017 to June 2022, and the data for monthly gas consumption used the technique of walk-forward validation (WFV) to evaluate the generalization ability prediction. Based on error scores, the study results compared the accuracy rates of the four models’ approaches show that: the Bi-GRU model was the best prediction for Bulgaria, Hungary, and Romania, the LSTM model best approached Slovakia, and the GRU model was the best suitable for Poland and Czech. The predicted methods for effectively managing the natural gas resource are crucial for enhancing gas consumption’s efficiency and reducing its impact on the environment. In addition, the result may contribute to stakeholders taking the right decision for energy planning. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Prediction models

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：