检索结果-内蒙古大学图书馆

Learning to Branch in Combinatorial Optimization With Graph Pointer Networks

IEEE/CAA Journal of Automatica Sinica 2024年第1期11卷 157-169页

作者： Rui Wang Zhiming Zhou Kaiwen Li Tao Zhang Ling Wang Xin Xu Xiangke Liao IEEE the Xiangjiang Laboratory and the College of Systems Engineering National University of Defense TechnologyChangsha 410073China the Institute of Automation Chinese Academy of SciencesBeijing 100190China the College of Systems Engineering National University of Defense TechnologyChangsha 410073 the Hunan Key Laboratory of Multi-Energy System Intelligent Interconnection Technology Changsha 410073China the Department of Automation Tsinghua UniversityBeijing 100084China the College of Intelligence Science and Technology National University of Defense TechnologyChangsha 410073China the College of Computer Science and Technology National University of Defense TechnologyChangsha 410073China

Traditional expert-designed branching rules in branch-and-bound(B&B) are static, often failing to adapt to diverse and evolving problem instances. Crafting these rules is labor-intensive, and may not scale well with complex *** the frequent need to solve varied combinatorial optimization problems, leveraging statistical learning to auto-tune B&B algorithms for specific problem classes becomes attractive. This paper proposes a graph pointer network model to learn the branch rules. Graph features, global features and historical features are designated to represent the solver state. The graph neural network processes graph features, while the pointer mechanism assimilates the global and historical features to finally determine the variable on which to branch. The model is trained to imitate the expert strong branching rule by a tailored top-k Kullback-Leibler divergence loss function. Experiments on a series of benchmark problems demonstrate that the proposed approach significantly outperforms the widely used expert-designed branching rules. It also outperforms state-of-the-art machine-learning-based branch-and-bound methods in terms of solving speed and search tree size on all the test instances. In addition, the model can generalize to unseen instances and scale to larger instances.

关键词： Branch-and-bound(B&B) combinatorial optimization deep learning graph neural network imitation learning

来源：评论

学校读者我要写书评

暂无评论

SketchFlex: Facilitating Spatial-Semantic Coherence in Text-to-Image Generation with Region-Based Sketches 25

SketchFlex: Facilitating Spatial-Semantic Coherence in Text-...

引用

2025 CHI Conference on Human Factors in Computing Systems, CHI 2025

作者： Lin, Haichuan Ye, Yilin Xia, Jiazhi Zeng, Wei Guangzhou China Academy of Interdisciplinary Studies The Hong Kong University of Science and Technology Guangdong China School of Computer Science and Engineering Central South University Changsha China Guangdong Guangzhou China The Hong Kong University of Science and Technology Hong Kong

ISBN: (纸本)9798400713941

Text-to-image models can generate visually appealing images from text descriptions. Efforts have been devoted to improving model controls with prompt tuning and spatial conditioning. However, our formative study highlights the challenges for non-expert users in crafting appropriate prompts and specifying fine-grained spatial conditions (e.g., depth or canny references) to generate semantically cohesive images, especially when multiple objects are involved. In response, we introduce SketchFlex, an interactive system designed to improve the flexibility of spatially conditioned image generation using rough region sketches. The system automatically infers user prompts with rational descriptions within a semantic space enriched by crowd-sourced object attributes and relationships. Additionally, SketchFlex refines users' rough sketches into canny-based shape anchors, ensuring the generation quality and alignment of user intentions. Experimental results demonstrate that SketchFlex achieves more cohesive image generations than end-to-end models, meanwhile significantly reducing cognitive load and better matching user intentions compared to region-based generation baseline. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Generative artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Brain Imaging and GRU-CNN Hybrid Framework for Alzheimer's Disease Progression Analysis 4

Brain Imaging and GRU-CNN Hybrid Framework for Alzheimer's D...

引用

4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Selvam, L. Dhamotharan, R. Krishnan, R. Santhana Francis Raj, J. Relin Sankar, G. Ram Yamini, G. Srm Madurai College for Engineering and Technology Department of Computer Science and Engineering Pottapalayam Sivaganga India Kalasalingam Academy of Research and Education Department of Computer Science and Information Technology Krishnankoil India Scad College of Engineering and Technology Department of Electronics and Communication Engineering Cheranmahadevi India Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Department of Electronics and Communication Engineering Chennai India Velammal College of Engineering and Technology Department of Computer Science and Engineering Madurai India Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Avadi Department of English Chennai India

ISBN: (纸本)9798331523923

Alzheimer's disease detection and classification using deep learning techniques offer significant advancements in early diagnosis and progression analysis of this neurodegenerative disorder. This approach leverages various medical imaging modalities such as MRI, PET, CT, and fMRI, alongside clinical data like cognitive test results and biomarkers, to assess brain changes associated with Alzheimer's. The focus is on classifying the disease into three stages: Mild, Moderate, and Severe, using advanced machine learning models. Deep learning techniques, such as EfficientNet B7 for 2D imaging and 3D CNNs for volumetric data, enable efficient feature extraction from medical scans. Additionally, Recurrent Neural Networks (RNNs) with GRU layers are used to capture temporal patterns from longitudinal data, improving predictive accuracy. Hybrid models, combining CNNs with attention mechanisms, further enhance classification. The model uses a multi-class classification approach with Softmax activation to predict the likelihood of each stage of Alzheimer's disease. Performance evaluation includes key metrics such as accuracy, precision, recall, F1 score, and AUC (Area Under the Curve). AUC can be calculated in two ways: individual AUC scores for each disease stage or a combined AUC for overall classification. This analysis provides a comprehensive understanding of the model's effectiveness in distinguishing between different stages of Alzheimer's. The results can be integrated into clinical workflows to assist healthcare professionals in diagnosing and monitoring Alzheimer's disease progression, offering a potential path toward more efficient and early intervention strategies. © 2025 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Hierarchical Loss for Bi-Level Classification of Speech into Language and Dialects

Hierarchical Loss for Bi-Level Classification of Speech into...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Angra, Ananya Muralikrishna, H. Dileep, A.D. Thenkanidiyoor, Veena MANAS Lab SCEE Indian Institute of Technology Mandi Mandi India Department of Electronics and Communication Engineering Manipal Institute of Technology Manipal Academy of Higher Education Manipal576104 India Department of Computer Science Engineering Indian Institute of Technology Dharwad Dharwad India Department of Computer Science Engineering National Institute of Technology Goa Goa India

ISBN: (纸本)9798350368741

Spoken dialect identification (DID) is a challenging task due to high similarities between the classes. This becomes further complicated when DID needs to be performed in a multilingual environment, especially when languages are from same language-family, as the scope for confusion is significantly high. In this paper, we propose multiple techniques to address this issue. These techniques motivate the DID system to respect the parent-child relationship between the languages and their dialects by performing bi-level classification of speech. Specifically, in our first approach, we train the language identification (LID) and DID models independently and then combine them together such that the output of LID model selects the DID model, which in turn decides the dialect. However, such independent training does not allow the model to learn the relations/similarities between dialects of different languages. To overcome this limitation, we propose to train an end-to-end DID model which is trained using dialects of all languages in the dataset. While such end-to-end training allows the model to learn inter-dialect similarities in a better way, it does not explicitly prevent a dialect from being misclassified into a dialect of a different language. To address this limitation, we propose a novel hierarchical loss, which motivates the model to maintain the parent-child relationship between language and dialects. Specifically, with the help of an auxiliary language classifier and a primary dialect classifier, the hierarchical loss penalizes the model heavily whenever parent-child relationship is not maintained, i.e., when predicted dialect does not belong to the predicted language. Experiments conducted on a set of closely related Indian languages shows that hierarchical loss based training leads to improvement in the performance. © 2025 IEEE.

关键词： Hierarchical loss Multi-lingual Spoken dialect identification

来源：评论

学校读者我要写书评

暂无评论

An Adaptive Reversible Data Hiding Scheme using Image Processing and Deep Learning Techniques 5

An Adaptive Reversible Data Hiding Scheme using Image Proces...

引用

5th International Conference on Sustainable Communication Networks and Application, ICSCNA 2024

作者： Allwyn, Philip Hemalatha, J. Sampathkumar, Nithyanantham Ramaraj, Kottaimalai Swain, Gandharba Sekar, M. AAA College of Engineering and Technology Department of Electronics and Communication Engineering Tamilnadu Sivakasi India AAA College of Engineering and Technology Department of Computer Science and Engineering Tamilnadu Sivakasi India Kalasalingam Academy of Research and Education Department of Computer Science and Engineering Tamilnadu Krishnankoil India Kalasalingam Academy of Research and Education Department of Electronics and Communication Engineering Tamilnadu Krishnankoil India Koneru Lakshmaiah Education Foundation Department of Computer Science and Engineering Andhrapradesh Vaddeswaram India AAA College of Engineering and Technology Department of Mechanical Engineering Tamilnadu Sivakasi India

ISBN: (纸本)9798331530013

Image steganography is the process of hiding data from digital photos without changing how they look on the outside. It functions as a covert mode of communication, giving access to confidential data without raising the suspicions of opponents or unauthorized parties. It provides an extra degree of secrecy and security for digital communications. Prediction-error expansion (PEE) is an extensively used method in the type of reversible data hiding (RDH) that hiding data with intact image quality in the prediction errors. High-accuracy prediction algorithms are typically essential for a PEE-based RDH implementation to be successful. Current PEE-based Reversible Data Hiding (RDH) systems frequently use a single forecasting approach, typically with worldwide optimization, but pay little attention to the local neighborhoods' pixel distribution features. This study aims to analyze the most recent RDH techniques and investigate image processing and deep learning (DL) techniques for reversible data hiding (DH) schemes. RDH is a difficult topic in general which has potential uses in today's digital world. Reversible data hiding techniques properly restore the cover medium while safely transferring secret data (SD). The goal of RDH techniques is to achieve both large capacity and tunable quality. Even though reversible data hiding has been the subject of indepth research recently, there hasn't yet been a thorough literature analysis to identify research gaps and potential future approaches. © 2024 IEEE.

关键词： Data Hiding Data Security Image Processing Pixel Sorting Reversible Data Hiding Steganography

来源：评论

学校读者我要写书评

暂无评论

Integrating Deep Q-Networks and YOLO with Autonomous Robots for Efficient Warehouse Management 5

Integrating Deep Q-Networks and YOLO with Autonomous Robots ...

引用

5th IEEE International Conference on Recent Trends in computer science and technology, ICRTCST 2024

作者： Vivek, N. Suriya, A. Kamala, V. Chinnasamy, N.V. Akshya, J. Sundarrajan, M. Choudhry, Mani Deepak SRM Institute of Science and Technology Department of Computational Intelligence Chennai India KGiSL Institute of Technology Department of Computer Science and Engineering Coimbatore India Karpagam Academy of Higher Education Department of Computer Science Coimbatore India SRM Institute of Science and Technology Department of Networking and Communications Chennai India KGiSL Institute of Technology Department of Information Technology Coimbatore India

ISBN: (纸本)9798350351378

Autonomous robotic cars' rapid, accurate, and efficient product handling and navigation are changing warehouse inventory management. These 24/7 smart vehicles save labor costs, human error, and operational inefficiencies. Time-consuming and incorrect manual scanning and recording is used for inventory management. Traditional autonomous car Q-learning algorithms may suffer in dynamic inventory circumstances and lack the complexity needed for real-time decision-making. Our research creates novel algorithms for autonomous robotic cars. The Deep Q-Network (DQN) we utilize for navigation adapts better to changing circumstances than ordinary Q-learning. YOLO object detection replaces slower QR code recognition for inventory monitoring. We use Decision Trees to find QR code data insights for decision-making. Our research results are promising. A greater cumulative reward curve over 100 simulated episodes gives DQN a 40% learning efficiency advantage over Q-learning. With 95% accuracy, YOLO outperformed Haar Cascade classifiers in QR code detection by 40% speed and 15% accuracy. Decision Trees identified QR code data 5.9% better than Rule-Based Systems, indicating inventory dependability. Autonomous robots can simplify inventory management in warehouses. Algorithmic technology allows these ambulances to drive intelligently, detect warehouse things faster and more accurately, and make decisions based on complicated data. These advanced artificial intelligence algorithms may make robotic systems more responsive and customized, changing inventory management subfields. These algorithms must be updated so autonomous cars can drive to meet logistical needs. © 2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Interplay between cerebral and muscular activations during motor tasks: An fNIRS-sEMG study 46

Interplay between cerebral and muscular activations during m...

引用

46th Annual International Conference of the IEEE engineering in Medicine and Biology Society, EMBC 2024

作者： Huang, Ruisen Shang, Wenze Lin, Yongji Liang, Fengyan Yin, Ming Lu, Xiao-Ping Wu, Xinyu Gao, Fei Chinese Academy of Sciences Institute of Advanced Integration Technology Shenzhen Institute of Advanced Technology Guangdong Shenzhen China Hainan University School of Biomedical Engineering Hainan China Macau University of Science and Technology School of Computer Science and Engineering China

ISBN: (纸本)9798350371499

Understanding the intricate relationship between brain activations and muscle activations during motor tasks is crucial for elucidating motor control mechanisms and designing effective rehabilitation strategies. In this study, we investigated this relationship using functional near-infrared spectroscopy (fNIRS) and surface electromyography (sEMG) technologies. Three healthy male subjects performed lower-limb motor tasks while their brain activations and muscle activations were recorded simultaneously. The fNIRS data were converted to optical densities and filtered by lowpass filter, detrending, and short-separation regression to remove artifacts. The filtered signals were fed to a general linear model with the desired hemodynamic responses (dHRFs) constructed with Balloon model and simulated physiological noises reconstructed with the parameters obtained from the spectrum analysis of baseline signals. Despite the extracted hemodynamic responses from fNIRS data, sEMG signals were lowpass and analyzed to detect muscle activations. Results revealed that cerebral activations during lateral leg flexion tasks were similar yet only partially consistent with those observed in stand-up tasks. Significant correlations between brain and muscle activations were also observed, highlighting the complex interplay between central and peripheral neural mechanisms during motor control. These findings provide valuable insights into the neural basis of motor control and have implications for developing personalized rehabilitation interventions and assistive technologies to enhance motor function in clinical populations. © 2024 IEEE.

关键词： Near infrared spectroscopy

来源：评论

学校读者我要写书评

暂无评论

KSKV:Key-Strategy for Key-Value Data Collection with Local Differential Privacy

引用

computer Modeling in engineering & sciences 2024年第6期139卷 3063-3083页

作者： Dan Zhao Yang You Chuanwen Luo Ting Chen Yang Liu Artificial Intelligence Development Research Center Institute of Scientific and Technical Information of ChinaBeijing100038China Industry Development Department NSFOCUS Inc.BeijingChina School of Information Science and Technology Beijing Forestry UniversityBeijing100083China School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu610054China Institute of Computing Technology China Academy of Railway Sciences Corporation LimitedBeijing10081China

In recent years,the research field of data collection under local differential privacy(LDP)has expanded its focus fromelementary data types to includemore complex structural data,such as set-value and graph ***,our comprehensive review of existing literature reveals that there needs to be more studies that engage with key-value data *** studies would simultaneously collect the frequencies of keys and the mean of values associated with each ***,the allocation of the privacy budget between the frequencies of keys and the means of values for each key does not yield an optimal utility *** the importance of obtaining accurate key frequencies and mean estimations for key-value data collection,this paper presents a novel framework:the Key-Strategy Framework forKey-ValueDataCollection under ***,theKey-StrategyUnary Encoding(KS-UE)strategy is proposed within non-interactive frameworks for the purpose of privacy budget allocation to achieve precise key frequencies;subsequently,the Key-Strategy Generalized Randomized Response(KS-GRR)strategy is introduced for interactive frameworks to enhance the efficiency of collecting frequent keys through group-anditeration *** strategies are adapted for scenarios in which users possess either a single or multiple key-value ***,we demonstrate that the variance of KS-UE is lower than that of existing *** claims are substantiated through extensive experimental evaluation on real-world datasets,confirming the effectiveness and efficiency of the KS-UE and KS-GRR strategies.

关键词： Key-value local differential privacy frequency estimation mean estimation data perturbation

来源：评论

学校读者我要写书评

暂无评论

Using transformers and Bi-LSTM with sentence embeddings for prediction of openness human personality trait

引用

PeerJ computer science 2025年 11卷 1-42页

作者： Naz, Anam Khan, Hikmat Ullah Alsahfi, Tariq Alhajlah, Mousa Alshemaimri, Bader Daud, Ali Department of Information Technology University of Sargodha Punjab Sargodha Pakistan Department of Information Systems and Technology College of Computer Science and Engineering University of Jeddah Jeddah Saudi Arabia Applied of Computer Science College King Saud University Riyadh Saudi Arabia Software Engineering Department College of Computing and Information Sciences King Saud University Riyadh Saudi Arabia Faculty of Resilience Rabdan Academy Abu Dhabi United Arab Emirates

Understanding human personality traits is significant as it helps in decision making related to consumers’ behavior, career counselling, team building and top candidates’ selection for recruitment. Among various traits, openness is essential as it shows both diverse aspects of sensitive nature or intuitive nature. The individuals having a sensing nature tends to be more practical and prefer to focus on concrete information whereas the users having intuitive trait type is characterized by a focus on abstract ideas, creative thinking and future-oriented perspectives. In this research work, we aim to explore diverse natural language processing (NLP) based features and apply state of the art deep learning algorithms for openness trait prediction. Using standard Myers-Briggs Type Indicator (MBTI) dataset, we propose the use of the latest deep features of sentence embeddings which captures contextual semantics of the content to be used with deep learning models. For comparison, we explore textual features of Frequency-Inverse Document (TF-IDF) and parts of speech (POS) tagging with machine learning models and deep features of word2vec and global vectors for word representation (GloVe) with deep learning models. The comprehensive empirical analysis reveals that TF-IDF used with gradient boosting achieves high accuracy of 90% whereas, the deep feature of sentence embeddings when used and with deep model bidirectional long short-term memory (Bi-LSTM) achieves 90.5% accuracy. The best results have been achieved using the latest Transformer-based DistilBERT, which achieves the highest accuracy of 92% outperforming the existing studies in relevant literature. Subjects Human-computer Interaction, Artificial Intelligence, Data Mining and Machine Learning, Natural Language and Speech © 2025 xxxx

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Job-SDF: A Multi-Granularity Dataset for Job Skill Demand Forecasting and Benchmarking 38

Job-SDF: A Multi-Granularity Dataset for Job Skill Demand Fo...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Chen, Xi Qin, Chuan Fang, Chuyu Wang, Chao Zhu, Chen Zhuang, Fuzhen Zhu, Hengshu Xiong, Hui University of Science and Technology of China China Computer Network Information Center Chinese Academy of Sciences China Baidu Inc. China Institute of Artificial Intelligence Beihang University China SKLSDE School of Computer Science Beihang University China AI Thrust The Hong Kong University of Science and Technology Guangzhou China Department of Computer Science and Engineering The HongKong University of Science and Technology Hong Kong

In a rapidly evolving job market, skill demand forecasting is crucial as it enables policymakers and businesses to anticipate and adapt to changes, ensuring that workforce skills align with market needs, thereby enhancing productivity and competitiveness. Additionally, by identifying emerging skill requirements, it directs individuals towards relevant training and education opportunities, promoting continuous self-learning and development. However, the absence of comprehensive datasets presents a significant challenge, impeding research and the advancement of this field. To bridge this gap, we present Job-SDF, a dataset designed to train and benchmark job-skill demand forecasting models. Based on millions of public job advertisements collected from online recruitment platforms, this dataset encompasses monthly recruitment demand. Our dataset uniquely enables evaluating skill demand forecasting models at various granularities, including occupation, company, and regional levels. We benchmark a range of models on this dataset, evaluating their performance in standard scenarios, in predictions focused on lower value ranges, and in the presence of structural breaks, providing new insights for further research. Our code and dataset are publicly accessible via the https://***/Job-SDF/benchmark. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

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：