检索结果-内蒙古大学图书馆

Proceedings of the 38th International Conference on Neural information processing Systems

作者： Benyuan Meng Qianqian Xu Zitai Wang Zhiyong Yang Xiaochun Cao Qingming Huang Institute of Information Engineering CAS and School of Cyber Security University of Chinese Academy of Sciences Key Lab. of Intelligent Information Processing Institute of Computing Technology CAS and Peng Cheng Laboratory Key Lab. of Intelligent Information Processing Institute of Computing Technology CAS School of Computer Science and Tech. University of Chinese Academy of Sciences School of Cyber Science and Tech. Shenzhen Campus of Sun Yat-sen University School of Computer Science and Tech. University of Chinese Academy of Sciences and Key Lab. of Intelligent Information Processing Institute of Computing Technology CAS and Key Laboratory of Big Data Mining and Knowledge Management CAS

ISBN: (纸本)9798331314385

Diffusion models are powerful generative models, and this capability can also be applied to discrimination. The inner activations of a pre-trained diffusion model can serve as features for discriminative tasks, namely, diffusion feature. We discover that diffusion feature has been hindered by a hidden yet universal phenomenon that we call content shift. To be specific, there are content differences between features and the input image, such as the exact shape of a certain object. We locate the cause of content shift as one inherent characteristic of diffusion models, which suggests the broad existence of this phenomenon in diffusion feature. Further empirical study also indicates that its negative impact is not negligible even when content shift is not visually perceivable. Hence, we propose to suppress content shift to enhance the overall quality of diffusion features. Specifically, content shift is related to the information drift during the process of recovering an image from the noisy input, pointing out the possibility of turning off-the-shelf generation techniques into tools for content shift suppression. We further propose a practical guideline named GATE to efficiently evaluate the potential benefit of a technique and provide an implementation of our methodology. Despite the simplicity, the proposed approach has achieved superior results on various tasks and datasets, validating its potential as a generic booster for diffusion features. Our code is available at https://***/Darkbblue/diffusion-content-shift.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Analysis on the Training Mode of Master Students in Artificial Intelligence Field for Electronic information Professional Degree—Take Hunan Normal University as an Example 8th

An Analysis on the Training Mode of Master Students in Artif...

引用

8th EAI International Conference on e-Learning, e-Education, and Online Training, eLEOT 2022

作者： Fu, Weina Liu, Shuai Institute of Information Science and Engineering Hunan Normal University Hunan Changsha410081 China Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Hunan Changsha410081 China

ISBN: (纸本)9783031211638

The ability of independent innovation in the field of artificial intelligence is a key element to occupy the commanding heights of future technology and talent competition in China. The cultivation of artificial intelligence talents, especially the cultivation high-end talent, it is essential to promote the development of artificial intelligence industry. At present, the market for artificial intelligence talents is in great demand. There is a shortage of high-end artificial intelligence talents, the weak strength of artificial intelligence teachers in universities, and the incomplete education system for talent training and many other issues. This paper combines some measures taken by Hunan Normal University for graduate students in the field of artificial intelligence in the degree of electronic information, the necessity and feasibility of artificial intelligence training senior talents are discussed for social needs in the field of artificial intelligence, The training goals and methods of postgraduate students in the field of artificial intelligence under the professional category of electronic information major are clarified, and the construction of artificial intelligence field is planned. © 2022, ICST institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Professional aspects

来源：评论

学校读者我要写书评

暂无评论

Bi-Temporal Feature Relational Distillation for On-Board Lightweight Change Detection in Remote Sensing Imagery

引用

IEEE Transactions on Aerospace and Electronic Systems 2025年

作者： Wang, Guoqing Chen, He Li, Jie Wang, Jue Liu, Wenchao Chen, Liang Beijing Institute of Technology National Key Laboratory of Science and Technology on Space-Born Intelligent Information Processing Beijing100081 China Shanghai Aerospace Electronic Technology Institute Shanghai201109 China

Deep learning models for remote sensing change detection (CD) require significant resources, challenging real-time applications on spaceborne devices. Knowledge distillation (KD) technology is a potential solution that balances model size and CD performance. However, existing KD methods struggle to effectively transfer the teacher's ability to respond to changed landcovers, and also face challenges in assisting the student model in detecting changes in landcovers with subtle visual characteristics. To overcome these limitations, a bi-temporal feature relational distillation (FRD) framework is proposed. The FRD framework includes two designed distillation components: bi-temporal feature distance distillation (BFDD) and contrastive cluster representation distillation (CCRD). BFDD guides the student model to learn the relationship between bi-temporal feature categories from the teacher model and introduces the iteration-wise random selection and bilinear interpolation strategy to solve feature mismatch problem between the teacher and the student. CCRD describes the pixel-level and clusterlevel feature distributions through the designed contrastive cluster (CC) module, guiding the student model to align the change/nonchange features with the teacher model, so as to alleviate the confusion of change/non-change features. Extensive experiments and analyses have confirmed that the proposed FRD framework can develop lightweight models capable of achieving performance comparable to large models, thereby enhancing suitability for realtime CD. © 1965-2011 IEEE.

关键词： Change detection

来源：评论

学校读者我要写书评

暂无评论

Unifying Sum Rate Analysis and Optimization of RIS-Assisted MU-MIMO Networks: An Operator-Valued Free Probability Approach

Unifying Sum Rate Analysis and Optimization of RIS-Assisted ...

引用

IEEE International Conference on Communications in China (ICCC)

作者： Siqiang Wang Zhong Zheng Jing Guo Zesong Fei School of Information and Electronics Beijing Institute of Technology Beijing China National Key Laboratory of Science and Technology on Space-Born Intelligent Information Processing

ISBN: (数字)9798350378412

ISBN: (纸本)9798350378429

The fundamental performance limit of wireless systems is closely related to the transmission strategy of the transceivers as well as the wireless channels. As the channels are determined by the intrinsic characteristics of the environments, traditional system design is transceiver-centric, which adapts the transmit signals to the environments. Sharply different from the transceiver-centric design, by introducing the reconfigurable intelligent surface (RIS) into wireless systems, both the transceivers and the channels can be designed by properly adjusting the deployment of RIS panels as well as the phase-change and amplification of the RIS elements. However, as the RIS induces structural changes to the wireless channels, it is difficult to analyze and optimize the RIS-assisted systems in a unified framework. In this paper, we adopt the operatorvalued free probability theory to analyze the ergodic sum rate of an uplink MU-MIMO communication systems assisted by multiple cascaded RIS panels, where the sum rate is derived for rather generic system configurations, allowing arbitrary number of transceivers’ antennas, number of RIS panels, and a generic fading model. The obtained sum rate expression also enables a unified optimization framework that jointly designs the multi-antenna transceivers and the RIS configurations. Finally, numerical results are provided to verify the accuracy of the derived ergodic sum rate, and to show the performance of the joint optimization algorithms.

关键词： Wireless communication Fading channels Accuracy Reconfigurable intelligent surfaces Approximation algorithms Transceivers Numerical models

来源：评论

学校读者我要写书评

暂无评论

Integrating Multi-scale Contextualized information for Byte-based Neural Machine Translation

arXiv

引用

arXiv 2024年

作者： Huang, Langlin Feng, Yang Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences China Key Laboratory of AI Safety Chinese Academy of Sciences China University of Chinese Academy of Sciences China

Subword tokenization is a common method for vocabulary building in Neural Machine Translation (NMT) models. However, increasingly complex tasks have revealed its disadvantages. First, a vocabulary cannot be modified once it is learned, making it hard to adapt to new words. Second, in multilingual translation, the imbalance in data volumes across different languages spreads to the vocabulary, exacerbating translations involving low-resource languages. While byte-based tokenization addresses these issues, byte-based models struggle with the low information density inherent in UTF-8 byte sequences. Previous works enhance token semantics through local contextualization but fail to select an appropriate contextualizing scope based on the input. Consequently, we propose the Multi-Scale Contextualization (MSC) method, which learns contextualized information of varying scales across different hidden state dimensions. It then leverages the attention module to dynamically integrate the multi-scale contextualized information. Experiments show that MSC significantly outperforms subword-based and other byte-based methods in both multilingual and out-of-domain scenarios. We have uploaded the code to github1 © 2024, CC BY-NC-ND.

关键词： Neural machine translation

来源：评论

学校读者我要写书评

暂无评论

Instance-Aware Diffusion Implicit Process for Box-Based Instance Segmentation 26

Instance-Aware Diffusion Implicit Process for Box-Based Inst...

引用

26th European Conference on Artificial Intelligence, ECAI 2023

作者： Ren, Hao Liu, Xingsong Huang, Junjian Wan, Ru Pu, Jian Lu, Hong Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Fudan University China Institute of Science and Technology for Brain-inspired Intelligence Fudan University China Mogo.ai Information and Technology Co. Ltd. Shanghai China

ISBN: (纸本)9781643684369

The diffusion model has demonstrated impressive performance in image generation, but its potential for discriminative tasks such as instance segmentation remains unexplored. In this paper, we propose an Instance-aware Diffusion Implicit Process (IDIP) framework for instance segmentation based on boxes. During training, IDIP diffuses ground-truth boxes across various time steps, extracting corresponding Region of Interest (RoI) features. Dynamic convolution is then used to predict boxes and categories for each RoI, and the mask head generates masks from these predictions. During inference, IDIP iteratively refines randomly generated boxes with the denoising diffusion implicit model, while the mask head derives final masks from RoIs based on the refined boxes. Our method surpasses existing approaches on the COCO benchmark, requiring fewer training steps and less memory resources due to its dynamic design and instance-aware characteristic. © 2023 The Authors.

关键词： Diffusion

来源：评论

学校读者我要写书评

暂无评论

HighlightRemover: Spatially Valid Pixel Learning for Image Specular Highlight Removal 24

HighlightRemover: Spatially Valid Pixel Learning for Image S...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Zhang, Ling Ma, Yidong Jiang, Zhi He, Weilei Bao, Zhongyun Fu, Gang Xu, Wenju Xiao, Chunxia School of Computer Science and Technology Hubei Key Laboratory of Intelligent Information Processing and Realtime Industrial Systems Wuhan University of Science and Technology Wuhan China School of Computer Science Wuhan University Wuhan China Department of Computing Hong Kong Polytechnic University Hong Kong Amazon Palo Alto United States

ISBN: (纸本)9798400706868

Recently, learning-based methods have made significant progress for image specular highlight removal. However, many of these approaches treat all the image pixels uniformly, overlooking the negative impact of invalid pixels on feature reconstruction. This oversight often leads to undesirable outcomes, such as color distortion or residual highlights. In this paper, we propose a novel image specular highlight removal network called HighlightRNet, which utilizes valid pixels as references to reconstruct the highlight-free image. To achieve this, we introduce a context-aware fusion block (CFBlock) that aggregates information in four directions, effectively capturing global contextual information. Additionally, we introduce a location-aware feature transformation module (LFTModule) to adaptively learn the valid pixels for feature reconstruction, thereby avoiding information errors caused by invalid pixels. With these modules, our method can produce high-quality highlight-free results without color distortion and highlight residual. Furthermore, we develop a multiple light image-capturing system to construct a large-scale highlight dataset called NSH, which exhibits minimal misalignment in image pairs and minimal brightness variation in non-highlight regions. Experimental results on various datasets demonstrate the superiority of our method over state-of-the-art methods, both qualitatively and quantitatively. © 2024 ACM.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

A Fuzzy Creative Works Generation Algorithm Based on Graph Neural Network 6th

A Fuzzy Creative Works Generation Algorithm Based on Graph N...

引用

6th Euro-China Conference on intelligent Data Analysis and Applications, ECC 2019

作者： Zhang, Fuquan Wang, Yiou Zhao, Guifen Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Minjiang University Fuzhou350121 China Beijing Institute of Science and Technology Information Beijing100044 China

ISBN: (纸本)9789811650352

A fuzzy creative works generation algorithm based on graph neural network is proposed. Firstly, the multi-label fuzzy creative data set is constructed. Secondly, fuzzy logical correlations between creative objects are dynamically calculated by graphical neural networks to capture relevant digital objects and their relationships. Thirdly, the projectiles, boundary markers and location words of the creative scene objects are generated by analyzing related attributes of each entity based on graph neural network. Fuzzy creative works are automatically generated by adjusting the appropriate location. Finally, the experiments are deployed, and experimental results show that the accuracy rate of the proposed algorithm is improved by 9.6% on average compared with the manual selection, which indicates the feasibility of the proposed algorithm replacing the manual selection. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Fuzzy neural networks

来源：评论

学校读者我要写书评

暂无评论

DCTracker: Differential Motion Compensation Based Multi-Object Tracking in Satellite Videos

DCTracker: Differential Motion Compensation Based Multi-Obje...

引用

Signal, information and Data processing (ICSIDP), IEEE International Conference on

作者： Yuting Shi Yin Zhuang He Chen Miaoxin Cai Jianlin Xie Shuyu Gan National Key Laboratory of Science and Technology on Space-Born Intelligent Information Processing (SBIIP) Beijing Institute of Technology Beijing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

Multi-object tracking (MOT) in satellite videos is a vital technique in Earth observation, serving various purposes such as smart city management and military target reconnaissance. However, satellite videos face several challenges, including wide observation ranges, low target pixel proportions, and significant background interference such as cloud drift and lighting changes. Relying solely on appearance features from single-frame videos is insufficient for fully identifying targets; extracting target motion information across multiple frames is also necessary for *** paper introduces an innovative integrated real-time architecture for multi-object tracking, named DCTracker. Specifically, to address the loss of target motion information and the issue of weak targets being easily obscured by noise, we designed a differential compensation module to enhance discrimination between targets and noise by analyzing the motion characteristics of targets across consecutive video frames. Experimental results show that DCTracker achieves advanced performance, with an accuracy score of 61.3% on the VISO dataset, demonstrating its effectiveness in handling complex tracking scenarios.

关键词： Target tracking Satellites Smart cities Noise Semantics Reconnaissance Feature extraction Spatiotemporal phenomena Remote sensing Videos

来源：评论

学校读者我要写书评

暂无评论

Interpretable Object Recognition by Semantic Prototype Analysis

Interpretable Object Recognition by Semantic Prototype Analy...

引用

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Qiyang Wan Ruiping Wang Xilin Chen Institute of Computing Technology CAS Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS) Beijing China University of Chinese Academy of Sciences Beijing China

People can usually give reasons for recognizing a particular object as a specific category, using various means such as body language (by pointing out) and natural language (by telling). This inspires us to develop a recognition model with such principles to explain the recognition process to enhance human trust. We propose Semantic Prototype Analysis Network (SPANet), an interpretable object recognition approach that enables models to explicate the decision process more lucidly and comprehensibly to humans by "pointing out where to focus" and "telling about why it is" simultaneously. With the proposed method, some part prototypes with semantic concepts will be provided to elaborate on the classification together with a group of visualized samples to achieve both part-wise and semantic interpretability. The results of extensive experiments demonstrate that SPANet is able to recognize objects almost as well as the non-interpretable models, at the same time generating intelligible explanations for its decision process.

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：