检索结果-内蒙古大学图书馆

Synaptic learning with augmented spikes

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Yu, Qiang Song, Shiming Ma, Chenxiang Pan, Linqiang Tan, Kay Chen Tianjin Key Laboratory of Cognitive Computing and Application College of Intelligence and Computing Tianjin University Tianjin China Key Laboratory of Image Information Processing and Intelligent Control Institute of Artificial Intelligence School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China The Department of Computer Science City University of Hong Kong Hong Kong

Traditional neuron models use analog values for information representation and computation, while all-or-nothing spikes are employed in the spiking ones. With a more brain-like processing paradigm, spiking neurons are more promising for improvements on efficiency and computational capability. They extend the computation of traditional neurons with an additional dimension of time carried by all-or-nothing spikes. Could one benefit from both the accuracy of analog values and the time-processing capability of spikes? In this paper, we introduce a concept of augmented spikes to carry complementary information with spike coefficients in addition to spike latencies. New augmented spiking neuron model and synaptic learning rules are proposed to process and learn patterns of augmented spikes. We provide systematic insight into the properties and characteristics of our methods, including classification of augmented spike patterns, learning capacity, construction of causality, feature detection, robustness and applicability to practical tasks such as acoustic and visual pattern recognition. The remarkable results highlight the effectiveness and potential merits of our methods. Importantly, our augmented approaches are versatile and can be easily generalized to other spike-based systems, contributing to a potential development for them including neuromorphic computing. Copyright © 2020, The Authors. All rights reserved.

关键词： Neurons

Towards Understanding and Boosting Adversarial Transferability from a Distribution Perspective

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhu, Yao Chen, Yuefeng Li, Xiaodan Chen, Kejiang He, Yuan Tian, Xiang Zheng, Bolun Chen, Yaowu Huang, Qingming The Zhejiang University Hangzhou310013 China The Security Department of Alibaba Group China The CAS Key Laboratory of Electro-Magnetic Space Information University of Science and Technology of China China The Hangzhou Dianzi University China Zhejiang Provincial Key Laboratory for Network Multimedia Technologies China The Zhejiang University Zhejiang Provincial Key Laboratory for Network Multimedia Technologies China The Zhejiang University Zhejiang University Embedded System Engineering Research Center Ministry of Education of China China The University of Chinese Academy of Sciences China The Key Laboratory of Intelligent Information Processing Chinese Academy of Sciences China

Transferable adversarial attacks against Deep neural networks (DNNs) have received broad attention in recent years. An adversarial example can be crafted by a surrogate model and then attack the unknown target model successfully, which brings a severe threat to DNNs. The exact underlying reasons for the transferability are still not completely understood. Previous work mostly explores the causes from the model perspective, e.g., decision boundary, model architecture, and model capacity. adversarial attacks against Deep neural networks (DNNs) have received broad attention in recent years. An adversarial example can be crafted by a surrogate model and then attack the unknown target model successfully, which brings a severe threat to DNNs. The exact underlying reasons for the transferability are still not completely understood. Previous work mostly explores the causes from the model perspective, e.g., decision boundary, model architecture, and model capacity. Here, we investigate the transferability from the data distribution perspective and hypothesize that pushing the image away from its original distribution can enhance the adversarial transferability. To be specific, moving the image out of its original distribution makes different models hardly classify the image correctly, which benefits the untargeted attack, and dragging the image into the target distribution misleads the models to classify the image as the target class, which benefits the targeted attack. Towards this end, we propose a novel method that crafts adversarial examples by manipulating the distribution of the image. We conduct comprehensive transferable attacks against multiple DNNs to demonstrate the effectiveness of the proposed method. Our method can significantly improve the transferability of the crafted attacks and achieves state-of-the-art performance in both untargeted and targeted scenarios, surpassing the previous best method by up to 40% in some cases. In summary, our work provides new in

关键词： Deep neural networks

Truncated Beam Sweeping for Spatial Covariance Matrix Reconstruction in Hybrid Massive MIMO

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Liu, Yinsheng Duan, Hongtao Liao, Xi State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University Beijing100044 China Frontiers Science Center for Smart High-spped Railway System China Beijing100037 China Key Laboratory of Complex Environmental Communications School of Communication and Information Engineering Chongqing University of Posts and Telecommunications Chongqing400065 China

Spatial covariance matrix (SCM) is essential in many applications of multi-antenna systems such as massive multiple-input multiple-output (MIMO). For massive MIMO operating at millimeter-wave bands, hybrid analog-digital structure has been adopted to reduce the cost of radio frequency (RF) chains. In this situation, signals received at the antennas are unavailable to the digital receiver, and as a consequence, traditional sample average approach cannot be used for SCM reconstruction in hybrid massive MIMO. To address this issue, beam sweeping algorithm (BSA), which can reconstruct SCM effectively in hybrid massive MIMO, has been proposed in our previous work. In this paper, a truncated BSA is further proposed for SCM reconstruction by taking into account the patterns of antenna elements in the array. Due to the directive antenna pattern, sweeping results corresponding to predetermined direction-of-angles (DOA) far from the normal direction are small and thus can be replaced by predetermined constants. At the cost of negligible performance reduction, SCM can be reconstructed efficiently by sweeping only the predetermined DOAs that are close to the normal direction. In this way, BSA can be conducted much faster than its traditional counterpart. Insightful analysis will be also included to show the impact of truncation on the performance. © 2022, CC BY.

关键词： Millimeter waves

GRIDLESS TOMOGRAPHIC SAR IMAGING BASED ON ACCELERATED ATOMIC NORM MINIMIZATION WITH EFFICIENCY

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gao, Silin Zhang, Zhe Zhang, Bingchen Wu, Yirong Aerospace Information Research Institute Chinese Academy of Sciences Beijing100094 China Key Laboratory of Technology in Geo-spatial Information Processing and Application System Chinese Academy of Sciences Beijing100190 China School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing100049 China Key Laboratory of Intelligent Aerospace Big Data Application Technology Suzhou215123 China Suzhou Aerospace Information Research Institute Suzhou215123 China

Synthetic aperture radar (SAR) tomography (TomoSAR) enables the reconstruction and three-dimensional (3D) localization of targets based on multiple two-dimensional (2D) observations of the same scene. The resolving along the elevation direction can be treated as a line spectrum estimation problem. However, traditional super-resolution spectrum estimation algorithms require multiple snapshots and uncorrelated targets. Meanwhile, as the most popular TomoSAR imaging method in modern years, compressed sensing (CS) based methods suffer from the gridding mismatch effect which markedly degrades the imaging performance. As a gridless CS approach, atomic norm minimization can avoid the gridding effect but requires enormous computing resources. Addressing the above issues, this paper proposes an improved fast ANM algorithm to TomoSAR elevation focusing by introducing the IVDST-ANM algorithm, which reduces the huge computational complexity of the conventional time-consuming semi-positive definite programming (SDP) by the iterative Vandermonde decomposition and shrinkage-thresholding (IVDST) approach, and retains the benefits of ANM in terms of gridless imaging and single snapshot recovery. We conducted experiments using simulated data to evaluate the performance of the proposed method, and reconstruction results of an urban area from the SARMV3D-Imaging 1.0 dataset are also presented. Copyright © 2022, The Authors. All rights reserved.

关键词： Compressed sensing

Design of achromaticsurface microstructure for near-eye display with diffractive waveguide

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Xiao, Jiasheng Liu, Juan Han, Jian Wang, Yongtian China Beijing Engineering Research Center for Mixed Reality and Advanced Display School of Optoelectronics Beijing Institute of Technology Beijing100081 China Key Laboratory of Information System and Technology Beijing Institute of Control and Electronic Technology Beijing100038 China

Dispersion problem has always constrained the development of see-Through near-eye displays with diffractive waveguide. Here, we propose a design method and optical systems for achromatic surface microstructure composed with a triple-carved-sub-grating. The triple-carved-sub-grating with specified period is designed based on rigorous coupled wave theory, corresponding to Red, Green and Blue (RGB) wavelengths, respectively, with diffraction efficiencies of ±1st orders reach 47.5% under TE polarization. The surface microstructure with period of 18.9 is realized by integrating the three sub-gratings together, and it is verified numerically that the diffractive angle of RGB wavelengths is 35° and the crosstalk is less than 4.5% under the normal incidence of TE polarization, which imply the surface microstructure diffracts the certain RGB wavelengths achromatically. Via proper duplication it could be utilized as the combining optics of diffractive-waveguide near-eye display and head-up display because of its light and compact features. m. Copyright © 2017, The Authors. All rights reserved.

关键词： Diffraction

Tactics2D: A Highly Modular and Extensible Simulator for Driving Decision-making

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Li, Yueyuan Zhang, Songan Jiang, Mingyang Chen, Xingyuan Qian, Yeqiang Wang, Chunxiang Yang, Ming The Department of Automation Shanghai Jiao Tong University Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China The Global Institute of Future Technology Shanghai Jiao Tong University Shanghai200240 China The University of Michigan-Shanghai Jiao Tong University Joint Institute Shanghai Jiao Tong University Shanghai200240 China

Simulation is a prospective method for generating diverse and realistic traffic scenarios to aid in the development of driving decision-making systems. However, existing simulators often fall short in diverse scenarios or interactive behavior models for traffic participants. This deficiency underscores the need for a flexible, reliable, user-friendly open-source simulator. Addressing this challenge, Tactics2D adopts a modular approach to traffic scenario construction, encompassing road elements, traffic regulations, behavior models, physics simulations for vehicles, and event detection mechanisms. By integrating numerous commonly utilized algorithms and configurations, Tactics2D empowers users to construct their driving scenarios effortlessly, just like assembling building blocks. Users can effectively evaluate the performance of driving decision-making models across various scenarios by leveraging both public datasets and user-collected real-world data. For access to the source code and community support, please visit the official GitHub page for Tactics2D at https://***/WoodOxen/Tactics2D. Copyright © 2023, The Authors. All rights reserved.

关键词： Decision making

Numerical detection of Gaussian entanglement and its application to the identification of bound entangled Gaussian states

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Ma, Shan Xue, Shibei Guo, Yu Shu, Chuan-Cun School of Automation Central South University Changsha410083 China Peng Cheng Laboratory Shenzhen518000 China Department of Automation Shanghai Jiao Tong University Shanghai200240 China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China Hunan Provincial Key Laboratory of Flexible Electronic Materials Genome Engineering School of Physics and Electronic Science Changsha University of Science and Technology Changsha410114 China Hunan Key Laboratory of Super-Microstructure and Ultrafast Process School of Physics and Electronics Central South University Changsha410083 China

We present a numerical method for solving the separability problem of Gaussian quantum states in continuous-variable quantum systems. We show that the separability problem can be cast as an equivalent problem of determining the feasibility of a set of linear matrix inequalities. Thus, it can be efficiently solved using existent numerical solvers. We apply this method to the identification of bound entangled Gaussian states. We show that the proposed method can be used to identify bound entangled Gaussian states that could be simple enough to be producible in quantum optics. Copyright © 2020, The Authors. All rights reserved.

关键词： Quantum optics

Image Manipulation Detection Using the Attention Mechanism and Faster R-CNN

学校读者我要写书评

暂无评论

IAENG International Journal of Computer Science

IAENG International Journal of Computer Science 2023年第4期50卷 1261页

作者： Tan, Kang Li, Linna Huang, Qiongdan School of Control Science and Engineering Xi’an University of Post and Telecommunications Xi’an710121 China Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing Xi’an University of Post and Telecommunications Xi’an710121 China School of Communication and Information Engineering Xi’an University of Post and Telecommunications Xi’an710121 China

With the advancement of image authoring software, the traces of image tampering operation have become increasingly difficult to detect. To enhance the performance of manipulation detection, we propose a novel end-to-end image tampering detection method using the attention mechanism and two-stream Faster R-CNN. In our model, we eliminate steganalysis rich model (SRM) filter, which suppresses image content, and employ a constrained convolution layer to adaptively extract image tamper features during model training. These features are then inputted into the backbone network in parallel with the RGB image. To emphasize image tamper features and restrict irrelevant features, we introduce the Convolutional Block Attention Module (CBAM) into the model, enabling better delivery of tamper-related information. Then, the output of backbone network is fed into the Region of Interest (RoI), where we replace max pooling operation with bilinear interpolation. This modification allows the model to retain more tamper information. After that, we use bilinear pooling for feature fusion of the two streams, and the fused results perform tamper classification operations. The RGB stream through Region Proposal Network (RPN) to achieve tamper localization. We evaluate our model on three publicly available standard image tampering datasets and demonstrate through experiments that our approach significantly improves the precision of manipulation detection and localization. © 2023, International Association of Engineers. All rights reserved.

关键词： Convolution

Research on Color Harmony Evaluation Method based on Color Complexity Measure

学校读者我要写书评

暂无评论

Research on Color Harmony Evaluation Method based on Color C...

Culture-oriented Science & Technology (ICCST), International Conference on

作者： Yanting Gao Jingjing Zhang Shuang Wang Yuncong Peng Aixia Tang School of Information and Communication Engineering Communication University of China Beijing China Ministry of Culture and Tourism Key Laboratory of Acoustic Visual Technology and Intelligent Control System (Communication University of China) Beijing Key Laboratory of Modern Entertainment Technology (Communication University of China) School of Computer Science and Cybersecurity Communication University of China Beijing China

ISBN: (数字)9781728181387

ISBN: (纸本)9781728181394

Color harmony is an important factor affecting the aesthetic feeling of images and plays an important role in visual perception. It can evaluate the feasibility of images by constructing the computational model of human aesthetic thinking. In this paper, a color harmony evaluation method based on color complexity was implemented. Given the classical color harmony model which is suitable for the simple color combination but not suitable for the complex image, and the local feature redundancy caused by grid sampling, an adaptive sampling based on CCM(color complexity measure) was implemented. Then, K-means was adopted to extract local color harmony features. Finally, a model was constructed between local color harmony features and the overall color harmony by SVM. The results show that the method implemented in this paper can represent the color harmony of complex images accurately and effectively.

关键词： Image color analysis Support vector machines Feature extraction Visual perception Wheels Computational modeling Complexity theory