检索结果-内蒙古大学图书馆

HRL-Painter: Optimal planning painter based on hierarchical reinforcement learning

NEUROCOMPUTING 2025年 636卷

作者： Zhang, Jiong Xu, Guangxin Zhang, Xiaoyan Shenzhen Univ Coll Comp Sci & Software Engn Shenzhen 518060 Guangdong Peoples R China

Stroke-based rendering method has shown its superiority in generating stylized paintings from realistic photographs. However, the existing methods often divide the image into regular blocks for parallel painting or start painting by progressively narrowing down the painting region from the entire canvas. Not only does this lead to an irrational allocation of stroke resources, but also deviates from the painting approach employed by human artists. To address this, we propose a novel painting method based on hierarchical reinforcement learning, namely HRL-Painter, which consists of a high-level agent that strategically plans the sequence of painting regions and a low-level agent that carries out specific painting tasks in the corresponding regions. In the initial stage, we consider the entire canvas as the painting region and then use a small number of strokes for a rough depiction. Next, our high-level agent plans the optimal sequence of painting regions based on the content of the target image, taking into account the error between the current canvas and the target image. Finally, the low-level agent is dedicated to executing detailed painting tasks within the painting regions proposed by the high-level agent. Extensive experiments on standard datasets including CelebA, ImageNet, CUB-200 Birds and Stanford Cars-196 demonstrate that our proposed hierarchical painting agent not only produce high-quality canvases but also exhibit a painting process that closely resembles the human painting style, showcasing excellent interpretability.

关键词： Optimal planning Hierarchical reinforcement learning Stroke-based rendering

来源：评论

学校读者我要写书评

暂无评论

Tracking Computer Vision Algorithm Based on Fusion Twin Network

引用

INTERNATIONAL JOURNAL OF ADVANCED COMPUTER science AND APPLICATIONS 2024年第10期15卷 921-931页

作者： Wang, Xin Hunan Coll Informat Software Inst Changsha 410016 Peoples R China

learning technology has promoted the rapid development of visual object tracking, among which algorithms based on twin networks are a hot research direction. Although this method has broad application prospects, its performance is often greatly reduced when encountering target occlusion or similar objects in the background. In response to this issue, a method is proposed to integrate channel and spatial dimension attention mechanisms into the backbone architecture of twin networks, to optimize the algorithm's recognition accuracy for tracking targets and its stability in changing environments. Then, a region recommendation network based on adaptive anchor box generation is adopted, combined with twin networks to enhance the network's modeling ability for complex situations. Finally, a new visual tracking algorithm is designed. Through comparative experiments, the success rate of the former increased by 0.6% and 0.9% respectively on the two datasets, and its accuracy also increased by 1.2% and 1.8% accordingly. The success rate of the latter increased by 1.5% and 1.2% respectively in the two datasets, and the accuracy also increased by 1.2% and 0.6% respectively. From this, the improved algorithm can improve the performance of target tracking and has certain application potential in visual target tracking.

关键词： Visual tracking twin network integration attention mechanism self-adaption

来源：评论

学校读者我要写书评

暂无评论

Multi-Resolution Expansion of Analysis in Time-Frequency Domain for Time Series Forecasting

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2024年第11期36卷 6667-6680页

作者： Yan, Kaiwen Long, Chen Wu, Huisi Wen, Zhenkun Shenzhen Univ Coll Comp Sci & Software Engn Shenzhen 518060 Peoples R China

Time series forecasting plays a crucial role in various real-world applications, such as finance, energy, traffic, and healthcare, providing valuable insights for decision-making processes. The aggregation of information windows with different resolutions has proven effective in time series forecasting tasks and provides the model diverse contextual information. As a result, the network can better capture and model the heterogeneity present in the data, thereby improving performance. However, most of the current work focuses on extracting multilevel-resolution information without considering the possibility that important information can be supplemented. Meanwhile, these methods also tend to ignore the effect of resolution on frequency. To address these challenges, we introduce the Time-Frequency Domain Multi-Resolution Expansion Network (TFMRN) for long-series forecasting using multi-resolution time-frequency data. The proposed TFMRN aims to expand the data in both the time and frequency domains, enabling the model to capture finer details that may not be evident in the original data. In addition, we also propose an Information Gating Unit (IGU) to enhance the selection and guidance of rich information from the expanded time-frequency multi-resolution data. Experimental results demonstrate that the proposed method yields better performance compared with the state-of-the-art methods in both univariate and multivariate time forecasting tasks.

关键词： Time series analysis Time-frequency analysis Forecasting Feature extraction Data mining Task analysis Market research forecasting multi-resolution expansion time and frequency domain time series analysis

来源：评论

学校读者我要写书评

暂无评论

FairCoRe: Fairness-Aware Recommendation Through Counterfactual Representation Learning

引用

IEEE Transactions on Knowledge and Data Engineering 2025年第7期37卷 4049-4062页

作者： Bin, Chenzhong Liu, Wenqiang Zhang, Feng Chang, Liang Gu, Tianlong Guilin University of Electronic Technology School of Computer Science and Information Security Guilin541004 China Guilin University of Electronic Technology School of Business Guilin541004 China Guilin University of Electronic Technology Guangxi Key Laboratory of Trusted Software Guilin541004 China Jinan University College of Cyber Security Engineering Research Center of Trustworthy AI Ministry of Education Guangzhou510632 China

Eliminating bias from data representations is crucial to ensure fairness in recommendation. Existing studies primarily focus on weakening the correlation between data representations and sensitive attributes, yet may inadvertently steer the user representations toward another potential bias direction of the target attribute. Furthermore, they often overlook the impact of user preferences on capturing sensitive information, incurring inadequate bias elimination. In this paper, we propose a Fair Counterfactual Representations (FairCoRe) learning framework, which aims to ensure the neutrality of representations among all bias directions. First, we intervene on sensitive attributes to construct a counterfactual scenario. Then, two opposing attribute prediction tasks are respectively performed in ground-truth and counterfactual scenarios to encode sensitive information along different bias directions. Second, we design a bias-aware enhancement learning method that quantifies the respective correlation of user preferences and sensitive attributes to enhance sensitive information encoding. Finally, we introduce two mutual information optimization methods that optimize the representations to capture users’ interests and disentangle sensitive factors. Moreover, we propose an attribute neutralization strategy that refines the learned representations, ensuring sensitive attribute neutrality. Extensive experiments demonstrate that our method achieves the optimal fairness and competitive accuracy compared to state-of-the-art methods. © 1989-2012 IEEE.

关键词： Accuracy Mutual Information Correlation Data Visualization Medical Services Training Measurement Disentangled Representation Learning Semantics Recommender Systems Recommender System Recommendation Fairness Representation Learning Counterfactual Representations Potential Bias Mutual Information User Preferences Sensitive Attributes Counterfactual Scenario User Representation Time Complexity Generative Adversarial Networks Learning Objectives Demographic Groups Baseline Methods Recommender Systems Minimum Method Prediction Loss Preference Information Graph Convolution User Side Recommendation Model Counterfactual Thinking Maximum Mutual Information Representation Learning Methods Maximization Method Recommendation Accuracy Recommendation Method Recommendation Task Notions Of Fairness Negative Samples Mutual Information Estimation

来源：评论

学校读者我要写书评

暂无评论

Superatomic-based chirality:Asymmetric structures constructed by superatoms

Aggregate

引用

Aggregate 2024年第4期5卷 220-225页

作者： Famin Yu Rui Li Xinrui Yang Yulei Shi Zhigang Wang Institute of Atomic and Molecular Physics Jilin UniversityChangchunChina Key Laboratory of Material Simulation Methods and Software of Ministry of Education College of PhysicsJilin UniversityChangchunChina Department of Physics Capital Normal UniversityBeijingChina

Chirality is one of the fundamental properties of molecules traditionally con-structed from ***,we report for thefirst time the successful construction of asymmetric chiral structures utilizing highly symmetric endohedral metallo-fullerene superatoms based on their own bonding ***,stable mirror-symmetric sinister and rectus structures are obtained by selecting a super-atom capable of forming four chemical bonds as the chiral *** analysis shows that the chiral vibration frequency of superatomic assemblies can be as low as a few wavenumbers,which greatly expands the range of chiral spectra com-pared to atom-based *** term this type of chirality based on superatoms as“superatomic-based chirality”.It is anticipated that this work will significantly expand the variety of chiral structures at the atomic level.

关键词： assembly atomic level chirality superatom

来源：评论

学校读者我要写书评

暂无评论

A Progressive Semantic-Aware Fusion Network for Remote Sensing Object Detection

引用

APPLIED scienceS-BASEL 2025年第8期15卷 4422-4422页

作者： Li, Lerong Wang, Jiayang Liao, Yue Qian, Wenbin Jiangxi Agr Univ Sch Software Nanchang 330045 Jiangxi Peoples R China

Object detection in remote sensing images has gained prominence alongside advancements in sensor technology and earth observation systems. Although current detection frameworks demonstrate remarkable achievements in natural imagery analysis, their performance degrades when applied to remote imaging scenarios due to two inherent limitations: (1) complex background interference, which causes object features to be easily obscured by noise, leading to reduced detection accuracy;(2) the variation in object scales leads to a decrease in the model's generalization ability. To address these issues, we propose a progressive semantic-aware fusion network (ProSAF-Net). First, we design a shallow detail aggregation module (SDAM), which adaptively integrates features across different channels and scales in the early Neck stage through dynamically adjusted fusion weights, fully exploiting shallow detail information to refine object edge and texture representation. Second, to effectively integrate shallow detail information and high-level semantic abstractions, we propose a deep semantic fusion module (DSFM), which employs a progressive feature fusion mechanism to incrementally integrate deep semantic information, strengthening the global representation of objects while effectively complementing the rich shallow details extracted by SDAM, enhancing the model's capability in distinguishing objects and refining spatial localization. Furthermore, we develop a spatial context-aware module (SCAM) to fully exploit both global and local contextual information, effectively distinguishing foreground from background and suppressing interference, thus improving detection robustness. Finally, we propose auxiliary dynamic loss (ADL), which adaptively adjusts loss weights based on object scales and utilizes supplementary anchor priors to expedite parameter convergence during coordinate regression, thereby improving the model's positioning accuracy for targets. Extensive experiments on the RSOD,

关键词： feature fusion contextual information remote sensing images object detection

来源：评论

学校读者我要写书评

暂无评论

Mitigating the impact of mislabeled data on deep predictive models: an empirical study of learning with noise approaches in software engineering tasks

引用

AUTOMATED software ENGINEERING 2024年第1期31卷 33-33页

作者： Shen, Jian Li, Zhong Lu, Yifei Pan, Minxue Li, Xuandong Nanjing Univ State Key Lab Novel Software Technol Nanjing Peoples R China

Deep predictive models have been widely employed in software engineering (SE) tasks due to their remarkable success in artificial intelligence (AI). Most of these models are trained in a supervised manner, and their performance heavily relies on the quality of training data. Unfortunately, mislabeling or label noise is a common issue in SE datasets, which can significantly affect the validity of models trained on such datasets. Although learning with noise approaches based on deep learning (DL) have been proposed to address the issue of mislabeling in AI datasets, the distinct characteristics of SE datasets in terms of size and data quality raise questions about the effectiveness of these approaches within the SE context. In this paper, we conduct a comprehensive study to understand how mislabeled samples exist in SE datasets, how they impact deep predictive models, and how well existing learning with noise approaches perform on SE datasets. Through an empirical evaluation on two representative datasets for the Bug Report Classification and software Defect Prediction tasks, our study reveals that learning with noise approaches have the potential to handle mislabeled samples in SE tasks, but their effectiveness is not always consistent. Our research shows that it is crucial to address mislabeled samples in SE tasks. To achieve this, it is essential to take into account the specific properties of the dataset to develop effective solutions. We also highlight the importance of addressing potential class distribution changes caused by mislabeled samples and present the limitations of existing approaches for addressing mislabeled samples. Therefore, we urge the development of more advanced techniques to improve the effectiveness and reliability of deep predictive models in SE tasks.

关键词： Empirical study Deep predictive model Learning from noisy labels Label noise

来源：评论

学校读者我要写书评

暂无评论

Towards Accurate Alzheimer's Disease Diagnosis: Integrating Focused Linear Attention in Deep Learning Frameworks 8

Towards Accurate Alzheimer's Disease Diagnosis: Integrating ...

引用

8th International Artificial Intelligence and Data Processing Symposium, IDAP 2024

作者： Sam, Francis Qin, Zhiguang Addo, Daniel Arhin, Joseph Roger Ayivi, Williams Kwabena, Sarpong Muoka, Gladys Wavinya University of Electronic Science and Technology of China School of Information and Software Engineering Chengdu610054 China University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China

ISBN: (纸本)9798331531492

The early stage and accurate diagnosis of Alzheimer's Disease (AD) in neuroimaging remains a significant challenge. We introduce an innovative deep learning framework that incorporates a Focused Linear Attention (FLA) module to enhance the diagnosis of AD through medical imaging in this study. Our approach the strength of linear and softmax attention processes to enable the model to detect structural abnormalities and subtle patterns in brain pictures indicating various phases of cognitive decline. Our model achieves an accuracy of 93.21 %, together with a sensitivity of 92.74 % and a precision of 94.01 % in differentiating between non-demented, mild to moderately demented, and badly demented individuals according to experimental data. Our framework enhances overall diagnostic performance in multiclass classification by 1.85 % when compared to traditional machine learning models and baseline deep learning architectures. The incorporation of FLA greatly enhances these improvements, transforming our model into a powerful resource for the early diagnosis and intervention in AD. The results highlight how advanced deep learning methods, particularly Focused Linear Attention, could transform the early detection of Alzheimer's Disease. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Cryptanalysis and Improvement of Several Identity-Based Authenticated and Pairing-Free Key Agreement Protocols for IoT Applications

引用

SENSORS 2024年第1期24卷 61页

作者： Sun, Haiyan Li, Chaoyang Zhang, Jianwei Liang, Shujun Huang, Wanwei Zhengzhou Univ Light Ind Coll Software Engn Zhengzhou 450001 Peoples R China

Internet of Things (IoT) applications have been increasingly developed. Authenticated key agreement (AKA) plays an essential role in secure communication in IoT applications. Without the PKI certificate and high time-complexity bilinear pairing operations, identity-based AKA (ID-AKA) protocols without pairings are more suitable for protecting the keys in IoT applications. In recent years, many pairing-free ID-AKA protocols have been proposed. Moreover, these protocols have some security flaws or relatively extensive computation and communication efficiency. Focusing on these problems, the security analyses of some recently proposed protocols have been provided first. We then proposed a family of eCK secure ID-AKA protocols without pairings to solve these security problems, which can be applied in IoT applications to guarantee communication security. Meanwhile, the security proofs of these proposed ID-AKA protocols are provided, which show they can hold provable eCK security. Some more efficient instantiations have been provided, which show the efficient performance of these proposed ID-AKA protocols. Moreover, comparisons with similar schemes have shown that these protocols have the least computation and communication efficiency at the same time.

关键词： AKA identity-based cryptography eCK security model attacks

来源：评论

学校读者我要写书评

暂无评论

3D Human Pose Estimation via Graph Extended Spatio-Temporal Convolutional Network 9

3D Human Pose Estimation via Graph Extended Spatio-Temporal ...

引用

9th International Conference on Virtual Reality, ICVR 2023

作者： Jia, Yanhui Fan, Wanshu Zhou, Dongsheng Zhang, Qiang Dalian University National and Local Joint Engineering Laboratory of Computer Aided Design School of Software Engineering Dalian China

ISBN: (纸本)9798350345810

3D human pose estimation is an important premise for human behavior analysis and understanding, which has a wide range of applications in intelligent transportation, human-computer interaction, and animation production. Most existing works focus on extracting the feature relationship between frames by combining spatio-temporal information to reduce the error of attitude reconstruction. However, the majority of them often suffer from insufficient joint correlation characteristics. To address this problem, we propose a Graph Expand Spatiotemporal Convolutional Network, named GESC-Net, to improve the limitation of extracting human spatial structure features. To better enrich the feature of extracting local information, we develop a learnable symmetric connection (LSC) block in the spatial structure. Moreover, a CbAttantion block is also designed to obtain a larger view of the acquisition of global structure and get more effective features. We evaluate our approach on two standard benchmark datasets: Human3.6M and HumanEva-I. The quantitative and qualitative evaluation results demonstrate that the GESC-Net can achieve better 3D human posture estimation than existing state-of-the-art methods. © 2023 IEEE.

关键词： Convolution

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：