检索结果-内蒙古大学图书馆

An HHL-based algorithm for computing hitting probabilities of quantum walks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Guan, Ji Wang, Qisheng Ying, Mingsheng State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing100190 China Department of Computer Science and Technology Tsinghua University Beijing100084 China Centre for Quantum Software and Information University of Technology SydneyNSW2007 Australia

We present a novel application of the HHL (Harrow-Hassidim-Lloyd) algorithm - a quantum algorithm solving systems of linear equations - in solving an open problem about quantum walks, namely computing hitting (or absorption) probabilities of a general (not only Hadamard) one-dimensional quantum walks with two absorbing boundaries. This is achieved by a simple observation that the problem of computing hitting probabilities of quantum walks can be reduced to inverting a matrix. Then a quantum algorithm with the HHL algorithm as a subroutine is developed for solving the problem, which is faster than the known classical algorithms by numerical experiments. Copyright © 2020, The Authors. All rights reserved.

关键词： Quantum theory

Foster noisy label learning by exploiting noise-induced distortion in foreground localization

学校读者我要写书评

暂无评论

Neural Networks 2025年 191卷

作者： Ang Chen Feng Xu Xin Lyu Tao Zeng Xin Li College of Computer Science and Software Engineering Hohai University Nanjing 211100 China School of Computer Engineering Jiangsu Ocean University Lianyungang 222005 China Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University Nanjing 211100 China

Large-scale, well-annotated datasets are crucial for training deep neural networks. However, the prevalence of noisy-labeled samples can cause irreversible impairment to the generalization of models. Existing approaches have attempted to mitigate the impact of noisy labels by exploiting the different loss or confidence distributions between clean and noisy data to detect and correct noisy labels. This paper investigates the noise-induced distorting effect on foreground localization by tracking the model’s spatial attention distribution on visual activation maps. We observe that in clean samples, highly responsive regions usually focus on label-relevant foreground regions, whereas in noisy data, the model accidentally attends to uninformative background regions or cluttered object edges due to interference from label noise. Inspired by the observations, we propose a novel two-stage foreground localization-augmented noisy label learning framework named FLSC to concurrently boost the accuracy of sample selection and label correction for robust training. Specifically, FLSC first quantifies noise-induced distortion in foreground localization to foster conventional loss-based selection criteria by calculating the information reduction when deriving the foreground images from the original images based on the attention distribution. Next, we propose a noise-adaptive adversarial erasing strategy, which suppresses background activation by imposing adaptive erasure regularization, to eliminate overfitting to noisy samples while enhancing the learning of robust representations. To the best of our knowledge, it is the first attempt to exploit localization quality evaluation based on feature activation to address the label noise problem. Extensive experiments on synthetic and real-world datasets validate the superior performance of FLSC compared to state-of-the-art methods.

关键词：

An attention-based unsupervised adversarial model for movie review spam detection

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Gao, Yuan Gong, Maoguo Xie, Yu Qin, A.K. The School of Electronic Engineering Key Laboratory of Intelligent Perception and Image Understanding Ministry of Education Xidian University Xi'anShanxi Province710071 China The department of Computer Science and Software Engineering Swinburne University of Technology Melbourne Australia

With the prevalence of the Internet, online reviews have become a valuable information resource for people. However, the authenticity of online reviews remains a concern, and deceptive reviews have become one of the most urgent network security problems to be solved. Review spams will mislead users into making suboptimal choices and inflict their trust in online reviews. Most existing research manually extracted features and labeled training samples, which are usually complicated and time-consuming. This paper focuses primarily on a neglected emerging domain - movie review, and develops a novel unsupervised spam detection model with an attention mechanism. By extracting the statistical features of reviews, it is revealed that users will express their sentiments on different aspects of movies in reviews. An attention mechanism is introduced in the review embedding, and the conditional generative adversarial network is exploited to learn users' review style for different genres of movies. The proposed model is evaluated on movie reviews crawled from Douban, a Chinese online community where people could express their feelings about movies. The experimental results demonstrate the superior performance of the proposed approach. Copyright © 2021, The Authors. All rights reserved.

关键词： Generative adversarial networks

Detail retention and enhancement for camouflaged object detection

学校读者我要写书评

暂无评论

Neurocomputing 2025年 648卷

作者： Shangjing Chen Feng Xu Tao Zeng Xin Li Jiale Chen Xin Lyu College of Computer Science and Software Engineering Hohai University Nanjing 211100 China School of Computer Engineering Jiangsu Ocean University Lianyungang 222005 China Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University Nanjing 211100 China

The objective of camouflaged object detection is to segment objects that share similar visual attributes with their backgrounds. Existing methods leverage the consistency and variability of the object to pinpoint its location; however, when presented with samples exhibiting intricate edges, the model tends to overlook critical details. Toward this issue, we analyzed the reasons from the perspective of processing basic features and found that roughly compressing the feature channel could lead to the loss of fine features at the initial stage of the pipeline. Therefore, we introduce a multi-scale feature preprocessing approach named Soft Channel Compression prior to feature fusion, which retains key cues by enhancing the visibility of fine-grained features. We then utilize self-attention to fuse the results of multi-scale feature sampling, synthesizing contextual information between extended details and the main object. Furthermore, we propose the Hierarchical Receptive Field Complementarity module, using a divide-and-conquer strategy to enhance discriminative features with different scales in object details. This facilitates the integration of information across various ranges by employing a hierarchical manner to expand the receptive field, thereby achieving complementary information between local areas. Extensive experimental results on benchmark datasets demonstrate that the proposed method outperforms other state-of-the-art methods.

关键词： Camouflaged object detection Edge detail Hierarchical manner Multi-scale feature Regional information complementarity

Interlocking strategy in metal/polymer composites for enhanced anti-corrosion and sustained EMI shielding performance

学校读者我要写书评

暂无评论

Progress in Organic Coatings 2025年 207卷

作者： Sun, Jiahao Gao, Zhenliang Chen, Qiang Li, Yang Liu, Chuntai Shen, Changyu Liu, Xianhu State Key Laboratory of Structural Analysis Optimization and CAE Software for Industrial Equipment National Engineering Research Center for Advanced Polymer Processing Technology Zhengzhou University Zhengzhou450002 China State Key Laboratory of Solidification Processing School of Materials Science and Engineering Northwestern Polytechnical University Xi'an710072 China

The integration of highly conductive fillers into rubber matrices presents a persistent corrosion challenge, particularly in marine environments where the stealth capabilities of weaponry and equipment are paramount. Drawing upon interface engineering principles, we propose an interlocking strategy to regulate multiple interfacial interactions (including dipole-dipole, electrostatic and chemical bonding) between methyl vinyl silicone rubber (VMQ) and metal particles. This approach facilitates molecular conformational transformation and ordering in VMQ composites, inducing stable dipole distribution at the interface and significantly enhancing vulcanization crosslinking and filler-matrix adhesion. The engineered VMQ/Ag@Al/MWCNTs composite demonstrates exceptional performance, maintaining EMI shielding effectiveness >100 dB after 240 h copper accelerated salt spray (CASS) test, while exhibiting superior mechanical properties (2.5 MPa breaking strength, 120 % elongation) compared to pure silicone rubber (0.43 MPa). Moreover, the composite resists molecular orientation complexity, interface instability, and performance degradation during extended environmental exposures, offering a novel multi-scale interface control method for advanced rubber matrix composites. © 2025 Elsevier B.V.

关键词： Bond strength (chemical)

Bi-classifier with neighborhood aggregation for unsupervised domain adaptation

学校读者我要写书评

暂无评论

Information sciences 2025年 718卷

作者： Lü, Shuai Zhang, Xinyu Li, Zongze Li, Jingyao Kang, Meng Key Laboratory of Symbolic Computation and Knowledge Engineering (Jilin University) Ministry of Education Changchun 130012 China College of Computer Science and Technology Jilin University Changchun 130012 China College of Software Jilin University Changchun 130012 China

Recently, adversarial learning with bi-classifier has been confirmed to be effective in unsupervised domain adaptation, which can effectively promote the alignment of different distributions across domains. Most previous methods with bi-classifier exploit the output disagreement of the bi-classifier to learn transferable representations. However, only considering the disagreement between two classifiers cannot guarantee that the model performs well in the target domain. The main reason is that the model does not consider the local structure of the target domain data. To address this problem, we propose a Bi-Classifier with Neighborhood Aggregation (BCNA). Specifically, we provide two pseudo-labels for a target sample based on the nearest neighbor aggregation strategy. Then, we propose a self-supervised loss with two pseudo-labels based on confidence weight. In addition, we also propose a novel consistency regularization loss for bi-classifier. Unlike previous methods with bi-classifier, we achieve the goal of minimizing the disagreement between two classifiers from the perspective of different views of the sample. Our method can increase the discriminability of the model while ensuring the feature transferability. Extensive experiments show that our method has achieved state-of-the-art performance on three datasets that are widely used in unsupervised domain adaptation. © 2025 Elsevier Inc.

关键词： Bi-classifier Consistency regularization Neighborhood aggregation Self-supervision Unsupervised domain adaptation

Sliding sequential CVAE with time variant socially-aware rethinking for trajectory prediction

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhou, Hao Ren, Dongchun Yang, Xu Fan, Mingyu Huang, Hai National Key Laboratory of Science and Technology of Underwater Vehicle Harbin Engineering University Harbin China State Key Laboratory of Management and Control for Complex System Institute of Automation Chinese Academy of Sciences Beijing China Research Center for Autonomous Vehicles Meituan Beijing China College of Computer Science Wenzhou University Wenzhou China National Key Laboratory of Science and Technology of Underwater Vehicle Harbin Engineering University Harbin China

Pedestrian trajectory prediction is a key technology in many applications such as video surveillance, social robot navigation, and autonomous driving, and significant progress has been made in this research topic. However, there remain two limitations of previous studies. First, with the continuation of time, the prediction error at each time step increases significantly, causing the final displacement error to be impossible to ignore. Second, the prediction results of multiple pedestrians might be impractical in the prediction horizon, i.e., the predicted trajectories might collide with each other. To overcome these limitations, this work proposes a novel trajectory prediction method called CSR, which consists of a cascaded conditional variational autoencoder (CVAE) module and a socially-aware regression module. The cascaded CVAE module first estimates the future trajectories in a sequential pattern. Specifically, each CVAE concatenates the past trajectories and the predicted points so far as the input and predicts the location at the following time step. Then, the socially-aware regression module generates offsets from the estimated future trajectories to produce the socially compliant final predictions, which are more reasonable and accurate results than the estimated trajectories. Moreover, considering the large model parameters of the cascaded CVAE module, a slide CVAE module is further exploited to improve the model efficiency using one shared CVAE, in a slidable manner. Experiments results demonstrate that the proposed method exhibits improvements over state-of-the-art method on the Stanford Drone Dataset (SDD) and ETH/UCY of approximately 38.0% and 22.2%, respectively. © 2021, CC BY.

关键词： Trajectories

SketchyCOCO: Image generation from freehand scene sketches

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Gao, Chengying Liu, Qi Xu, Qi Wang, Limin Liu, Jianzhuang Zou, Changqing School of Data and Computer Science Sun Yat-sen University China State Key Laboratory for Novel Software Technology Nanjing University China Noah’s Ark Lab Huawei Technologies HMI Lab Huawei Technologies

We introduce the first method for automatic image generation from scene-level freehand sketches. Our model allows for controllable image generation by specifying the synthesis goal via freehand sketches. The key contribution is an attribute vector bridged Generative Adversarial Network called EdgeGAN, which supports high visual-quality object-level image content generation without using freehand sketches as training data. We have built a large-scale composite dataset called SketchyCOCO to support and evaluate the solution. We validate our approach on the tasks of both object-level and scene-level image generation on SketchyCOCO. Through quantitative, qualitative results, human evaluation and ablation studies, we demonstrate the method’s capacity to generate realistic complex scene-level images from various freehand sketches. Copyright © 2020, The Authors. All rights reserved.

关键词： Generative adversarial networks

Opportunistic Admissibility and Resource Allocation For Slicing-based Radio Access Networks

学校读者我要写书评

暂无评论

Opportunistic Admissibility and Resource Allocation For Slic...

International Conference on computer and Communications (ICCC)

作者： Zihao Qi Bin Cao Long Zhang Wei Hong Zhongyuan Zhao Jie Tong Xiang Bai China Electric Power Research Institute Beijing University of Posts and Telecommunications Beijing China State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China National Key Laboratory of Science and Technology on Communication University of Electronic Science and Technology of China Beijing Xiaomi Mobile Software Beijing China China Electric Power Research Institute Beijing China China Electronic Technology Cyber Security Co. Ltd. Beijing China

ISBN: (纸本)9781665409513

Network slicing (NS) is envisioned as a promising technology to meet the extremely diversified service requirements of users for future mobile networks. In radio access networks (RAN) slicing, the service providers (SPs) can rent network slicing instances from the infrastructure provider (InP) to meet the requirements of network services. However, both SPs and InP face the challenges of maintaining the quality of user experience and high profit in a dynamic environment, arising from random arrivals and departures of slice requests, uncertain resource availability, and multi-dimensional resource allocation. Therefore, admissibility and resource allocation become more complicated than that in traditional mobile networks. This paper proposes an opportunistic admissibility and resource allocation (OAR) policy to deal with the above challenges. To cope with the randomness of slice requests and resource availability, we first formulate this issue as a Markov Decision Process (MDP) problem to obtain the optimal admissibility strategy while maximizing the overall reward. Furthermore, we adopt a buyer-seller game-theoretic approach to optimize resource allocation, motivating SPs and InP to maximize their rewards. Numerical results show that the proposed OAR policy makes reasonable decisions effectively and steadily and outperforms the baseline scheme for system reward.

关键词： Network slicing Simulation Games Markov processes User experience Resource management III-V semiconductor materials