检索结果-内蒙古大学图书馆

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Wenju Sun Qingyong Li Jing Zhang Wen Wang Yangli-Ao Geng Beijing Key Lab of Traffic Data Analysis and Mining Beijing Jiaotong University Frontiers Science Center for Smart High-speed Railway System Beijing Jiaotong University The Key Laboratory of Cognitive Computing and Intelligent Information Processing of Fujian Education Institutions Wuyi University

The dilemma between plasticity and stability arises as a common challenge for incremental learning. In contrast, the human memory system is able to remedy this dilemma owing to its multilevel memory structure, which motivates us to propose a Bilevel Memory system with Knowledge Projection (BMKP) for incremental learning. BMKP decouples the functions of learning and remembering via a bilevel-memory design: a working memory responsible for adaptively model learning, to ensure plasticity; a long-term memory in charge of enduringly storing the knowledge incorporated within the learned model, to guarantee stability. However, an emerging issue is how to extract the learned knowledge from the working memory and assimilate it into the long-term memory. To approach this issue, we reveal that the parameters learned by the working memory are actually residing in a redundant high-dimensional space, and the knowledge incorporated in the model can have a quite compact representation under a group of pattern basis shared by all incremental learning tasks. Therefore, we propose a knowledge projection process to adaptively maintain the shared basis, with which the loosely organized model knowledge of working memory is projected into the compact representation to be remembered in the long-term memory. We evaluate BMKP on CIFAR-10, CIFAR-100, and Tiny-ImageNet. The experimental results show that BMKP achieves state-of-the-art performance with lower memory usage 1 1 The code is available at https://***/SunWenJu123/BMKP.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Uncovered Neurons Information-Based Fuzzing Method for DNN

An Uncovered Neurons Information-Based Fuzzing Method for DN...

引用

International Conference on Computer Supported Cooperative Work in Design

作者： Kecheng Tang Jian-Tao Zhou Xu Guo College of Computer Science Inner Mongolia University Hohhot China National & Local Joint Engineering Research Center of Intelligent Information Processing Technology for Mongolian Hohhot China Engineering Research Center of Ecological Big Data Ministry of Education Hohhot China Inner Mongolia Engineering Laboratory for Cloud Computing and Service Software Hohhot China Inner Mongolia Key Laboratory of Social Computing and Data Processing Hohhot China Inner Mongolia Key Laboratory of Discipline Inspection and Supervision Big Data Hohhot China Inner Mongolia Engineering Laboratory for Big Data Analysis Technology Hohhot China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

Fuzzing is increasingly being utilized as a method to test the reliability of Deep Learning (DL) systems. In order to detect more errors in DL systems, exploring the internal logic of more DNNs has become the main objective of fuzzing. Despite advancements in the seed selection aspect of fuzzing, considerable opportunities still exist for improving testing efficiency. Current research has issues with the repeated consideration of neurons in the model that will be covered in the future by other seeds, leading to redundant seeds and lower testing efficiency. Additionally, there is a lack of a method to measure the potential of seeds to increase coverage, making it difficult to select the most worthy seeds for mutation in each iteration. We propose an uncovered neurons information based (UNIB) fuzzing method for DNN. UNIB uses clustering methods to organize the seed queue based on initial seed data, aiming to enhance the coverage rate improved in each iteration. It also integrates coverage information from the testing phase to identify the seeds with the greatest potential. The experimental results show that UNIB achieved a higher NC than the second-best method by 1.1% and 3% in LetNet-4 and LetNet-5, respectively. UNIB consistently required the fewest number of iterations to reach the same NC as other methods. For both LetNet-4 and LetNet-5, the adversarial test case sets generated by UNIB exhibited the highest diversity.

关键词： Deep learning Federated learning Current measurement Neurons Neural networks Fuzzing Software

来源：评论

学校读者我要写书评

暂无评论

ULA Fitting for MIMO Radar

TechRxiv

引用

TechRxiv 2022年

作者： Shi, Wanlu Liu, Xiaoguang Li, Yingsong College of Information and Communication Engineering Harbin Engineering University Harbin150001 China School of Microelectronics Southern University of Science and Technology Shenzhen China Key Laboratory of Intelligent Computing Signal Processing Ministry of Education Anhui University Anhui Hefei China

Uniform linear array (ULA) fitting (UF) principle is a newly proposed sparse array (SA) design scheme that aims to design SAs using cascaded ULAs, based on which a number of SAs with desired capacity are devised. In this letter, we aim to introduce the UF idea to devise sparse MIMO arrays with reduced mutual coupling and increased uniform degrees of freedom. However, existing sparse MIMO array design approach needs the transmit and receive arrays in MIMO radar have hole-free difference coarrays (DCAs), which is no longer the best choice for UF principle due to the fact that SAs designed via UF cannot guarantee hole-free DCAs. In this regard, considering the difference coarray of the sum coarray (DCSC) concept, a novel sparse MIMO array design strategy which is suitable for UF principle is proposed. Based on the novel strategy and UF principle, a new sparse MIMO array is designed. Numerical simulations verify the good performance of the proposed sparse MIMO array in high coupling scenarios. © 2022, CC BY.

关键词： MIMO radar

来源：评论

学校读者我要写书评

暂无评论

A Noise Extraction Method for Cryo-EM Single-Particle Denoising

引用

Journal on Big Data 2022年第1期4卷 61-76页

作者： Huanrong Tang Sihan Wang Jianquan Ouyang Tianming Liu Key Laboratory of Intelligent Computing Information Processing Ministry of EducationSchool of Computer Science and School of Cyberspace ScienceXiangtan UniversityXiangtan411100China Department of Computer Science The University of GeorgiaGeorgia30301USA

Cryo-Electron Microscopy(cryo-EM)has become a powerful method to study the structure and function of biological ***,in clustering tasks based on the projection angle of particles in cryoEM,the noise considerably affects the clustering *** denoising algorithms are ineffective due to the extremely low signal-to-noise ratio(SNR)of cryo-EM images and the complexity of noise *** noise of a single particle greatly influences the orientation estimation of the subsequent clustering task,and the result of the clustering task directly affects the accuracy of the 3D *** this paper,we propose a construction method of cryo-EM denoising dataset that uses U-Net to extract noise blocks from cryoEM images,superimpose the noise block with the projected pure particles to construct our simulated *** we adopt a supervised generative adversarial network(GAN)with perceptual loss to train on our simulated dataset and denoise the real cryo-EM single *** method can solve the problem of poor denoising performance caused by assuming that the noise of the Gaussian distribution does not conform to the noise distribution of cryo-EM,and it can retain the useful information of particles to a great *** compared traditional image filtering methods and the classic deep learning denoising algorithm DnCNN on the simulated and real *** results show that the method based on deep learning has more advantages than traditional image denoising *** is worth mentioning that our method achieves a competitive peak signal to noise ratio(PSNR)and structural similarity(SSIM).Moreover,visualization results,indicate that our method can retain the structure information and orientation information of particles to a greater extent compared with other state-of-the-art image denoising *** means that our denoising task can provide considerable help for subsequent cryo-EM clustering tasks.

关键词： Cryo-EM noise extraction denoising GAN

来源：评论

学校读者我要写书评

暂无评论

Look Inside for More: Internal Spatial Modality Perception for 3D Anomaly Detection 39

Look Inside for More: Internal Spatial Modality Perception f...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Liang, Hanzhe Xie, Guoyang Hou, Chengbin Wang, Bingshu Gao, Can Wang, Jinbao College of Computer Science and Software Engineering Shenzhen University Shenzhen China Shenzhen Audencia Financial Technology Institute Shenzhen University Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China Department of Intelligent Manufacturing CATL Ningde China School of Computing and Artificial Intelligence Fuyao University of Science and Technology Fuzhou China School of Software Northwestern Polytechnical University Xi’an China Guangdong Provincial Key Laboratory of Intelligent Information Processing Shenzhen China

ISBN: (纸本)157735897X

3D anomaly detection has recently become a significant focus in computer vision. Several advanced methods have achieved satisfying anomaly detection performance. However, they typically concentrate on the external structure of 3D samples and struggle to leverage the internal information embedded within samples. Inspired by the basic intuition of why not look inside for more, we introduce a straightforward method named Internal Spatial Modality Perception (ISMP) to explore the feature representation from internal views fully. Specifically, our proposed ISMP consists of a critical perception module, Spatial Insight Engine (SIE), which abstracts complex internal information of point clouds into essential global features. Besides, to better align structural information with point data, we propose an enhanced key point feature extraction module for amplifying spatial structure feature representation. Simultaneously, a novel feature filtering module is incorporated to reduce noise and redundant features for further aligning precise spatial structure. Extensive experiments validate the effectiveness of our proposed method, achieving object-level and pixel-level AUROC improvements of 4.2% and 13.1%, respectively, on the Real3D-AD benchmarks. Note that the strong generalization ability of SIE has been theoretically proven and is verified in both classification and segmentation tasks. © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Improving Neural Radiance Fields with Depth-aware Optimization for Novel View Synthesis

arXiv

引用

arXiv 2023年

作者： Chen, Shu Li, Junyao Zhang, Yang Zou, Beiji The School of Computer Science School of Cyberspace Security Xiangtan University Xiangtan China Key Laboratory of Intelligent Computing & Information Processing Ministry of Education Xiangtan China School of Computer Science and engineering Central South University Changsha China

With dense inputs, Neural Radiance Fields (NeRF) is able to render photo-realistic novel views under static conditions. Although the synthesis quality is excellent, existing NeRF-based methods fail to obtain moderate three-dimensional (3D) structures. The novel view synthesis quality drops dramatically given sparse input due to the implicitly reconstructed inaccurate 3D-scene structure. We propose SfMNeRF, a method to better synthesize novel views as well as reconstruct the 3D-scene geometry. SfMNeRF leverages the knowledge from the self-supervised depth estimation methods to constrain the 3D-scene geometry during view synthesis training. Specifically, SfMNeRF employs the epipolar, photometric consistency, depth smoothness, and position-of-matches constraints to explicitly reconstruct the 3D-scene structure. Through these explicit constraints and the implicit constraint from NeRF, our method improves the view synthesis as well as the 3D-scene geometry performance of NeRF at the same time. In addition, SfMNeRF synthesizes novel sub-pixels in which the ground truth is obtained by image interpolation. This strategy enables SfMNeRF to include more samples to improve generalization performance. Experiments on three public datasets demonstrate that SfMNeRF surpasses state-of-the-art approaches given the spare inputs. Code is available at https://***/XTU-PR-LAB/SfMNeRF. Copyright © 2023, The Authors. All rights reserved.

关键词： Optical flows

来源：评论

学校读者我要写书评

暂无评论

Staggered Grid Scheme for the FFT-Based Methods

引用

Chinese Journal of Electronics 2019年第5期28卷 1066-1072页

作者： XIE Jiaye KONG Weibin PANG Lili SONG Weiju HUANG Zhixiang WU Xianliang Industrial Center Nanjing Institute of Technology State Key Laboratory of Millimeter Waves College of Information Engineering Yancheng Institute of Technology Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Anhui University

A staggered grid scheme is proposed to reduce both the total memory requirement and the CPU time of generating the corrected near matrix in the FFTbased methods. Two sets of Cartesian grids are used to project the source points and the field points, *** proposed method does not lower the efficiency of computing far matrix-vector products, compared with the traditional uniform Cartesian grid scheme. Some numerical experiments are provided to demonstrate both the correctness and the efficiency of the proposed method.

关键词： Green’s function Singularity RFG-FFT Staggered grid scheme Corrected near matrix

来源：评论

学校读者我要写书评

暂无评论

SFTA: Spiking Neural Networks Vulnerable to Spiking Feature Transferable Attack

SFTA: Spiking Neural Networks Vulnerable to Spiking Feature ...

引用

IEEE International Symposium on Parallel and Distributed processing with Applications and IEEE International Conference on Ubiquitous computing and Communications (ISPA/IUCC)

作者： Xuanwei Lin Chen Dong Ximeng Liu College of Computer and Data Science Fuzhou University Fujian Key Laboratory of Network Computing and Intelligent Information Processing (Fuzhou University) Key Lab of Information Security of Network Systems (Fuzhou University)

Many recent works have shown that spiking neural networks (SNNs) are vulnerable to well-crafted adversarial attacks in the white-box setting, which may lead to misclassification of the trained model. However, the information about the model is in a black-box setting and is usually unknown to the attacker, e.g., the parameters and structure of the model. So it is more challenging to counter the attack in the black-box setting, and there is still little research on this aspect in SNNs. Therefore, this work describes a transfer-based black box adversarial attack against spike feature space representations, providing insights for robust spike features in SNNs models. The attack is explicitly designed to be transferable, and more robust feature gradients are obtained by accumulating the feature gradients of the intermediate layers. Then we can successfully attack the black-box model by suppressing robust positive features and enhancing negative features. Extensive experimental evaluation demonstrates the proposed method’s effectiveness and superior performance on static and neuromorphic datasets. i.e., The attack success is 89.72% and 90% on the static MNIST and CIFAR10 datasets, and 85.97% and 69.09% on the neuromorphic N-MNIST and CIFARO-DVS datasets.

关键词： Correlation Neuromorphics Neural networks Closed box Ubiquitous computing Optimization Glass box

来源：评论

学校读者我要写书评

暂无评论

Vision-Based Fruit Recognition Via Multi-Scale Attention Cnn

SSRN

引用

SSRN 2022年

作者： Min, Weiqing Wang, Zhiling Yang, Jiahao Liu, Chunlin Jiang, Shuqiang The Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China University of Chinese Academy of Sciences Beijing100049 China

Fruit quality assessment, grading and sorting are of vital importance to fruit processing, and all these involve fruit recognition. Vision-based fruit recognition can recognize fruit automatically and further support more applications such as fruit-picking robots, self-checkout service, and dietary guidance. Recent works resort to Convolutional Neural Networks (CNNs) for vision-based fruit recognition because of their powerful expressive capacity. However, most works simply utilize existing CNNs and ignore characteristics of fruit images, resulting in sub-optimal recognition performance. To solve this problem, we adopt a Multi-Scale Attention Network (MSANet), which explores attention from different layers of CNNs and aggregates various visual attentional features from different levels into final comprehensive representations. Extensive evaluations on four fruit benchmark datasets demonstrate that the method achieves state-of-the-art recognition performance. In addition, we report various experimental results from different deep networks as baselines. Codes and models will be released upon publication. Models and code are available at http://123.57.42.89/codes/***. © 2022, The Authors. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Question generation based on chat-response conversion

Question generation based on chat-response conversion

引用

作者： Zhong, Sheng-Hua Peng, Jianfeng Liu, Peiqi College of Computer Science and Software Engineering Shenzhen University Shenzhen China The National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen China

Today, thanks to the major breakthrough of sequences to sequences model in the field of natural language, most of the dialogue generation tasks are focused on generating more effective responses. However, the responses proposed by the chat-bot are only a passive answer or assentation, which does not arouse the desire of people to continue communicating. How to transform the chat robot from a passive reply to an active questioner has become an urgent problem. In this paper, a question generalization method with four types of question proposing schemes are designed, implemented, and tested to automate question generation process. The proposed system is controlled by a probability-triggered multiple conversion mechanism to actively propose different types of questions. We embed our methods in the mainstream dialogue generation model and demonstrate its effectiveness in dialogue response generalization on a standard dataset. In addition, it achieves good performance in subjective conversational assessment. © 2019 John Wiley & Sons, Ltd.

关键词： dialogue response generalization system probability-triggered question generation

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：