检索结果-南通市图书馆

FSCIL-EACA: Few-Shot Class-Incremental Learning Network Based on Embedding Augmentation and Classifier Adaptation for Image Classification

引用

Chinese Journal of Electronics 2024年第1期33卷 139-152页

作者： Ruru ZHANG Haihong E Meina SONG School of Computer Science Beijing University of Posts and Telecommunications Education Department Information Network Engineering Research Center Beijing University of Posts and Telecommunications

The ability to learn incrementally is critical to the long-term operation of AI systems. Benefiting from the power of few-shot class-incremental learning(FSCIL), deep learning models can continuously recognize new classes with only a few samples. The difficulty is that limited instances of new classes will lead to overfitting and exacerbate the catastrophic forgetting of the old classes. Most previous works alleviate the above problems by imposing strong constraints on the model structure or parameters, but ignoring embedding network transferability and classifier adaptation(CA), failing to guarantee the efficient utilization of visual features and establishing relationships between old and new classes. In this paper, we propose a simple and novel approach from two perspectives: embedding bias and classifier bias. The method learns an embedding augmented(EA) network with cross-class transfer and class-specific discriminative abilities based on self-supervised learning and modulated attention to alleviate embedding bias. Based on the adaptive incremental classifier learning scheme to realize incremental learning capability,guiding the adaptive update of prototypes and feature embeddings to alleviate classifier bias. We conduct extensive experiments on two popular natural image datasets and two medical datasets. The experiments show that our method is significantly better than the baseline and achieves state-of-the-art results.

关键词： Training Adaptation models Visualization Image recognition Prototypes Self-supervised learning Power capacitors

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Ultralow-phase-noise and broadband frequency-hopping coupled optoelectronic oscillator under quiet point operation

引用

Photonics Research 2024年第8期12卷 1785-1793页

作者： HUI LIU MINGYANG GUO TIAN ZHANG JIAN DAI KUN XU State Key Laboratory of Information Photonics and Optical Communications Beijing University of Posts and Telecommunications School of Electronic Engineering Beijing University of Posts and Telecommunications

Advancements in microwave photonics have yielded novel approaches for generating high-purity microwave sources. Among these, optoelectronic oscillators（OEOs） and coupled optoelectronic oscillators（COEOs） have demonstrated the capability to generate frequency-independent microwaves with exceptionally low phase ***, the tunability of the oscillators is rather limited due to the necessity for narrowband electronic bandpass filters, presenting challenges in achieving both wide and rapid tuning capabilities. Here, we present a COEO featuring ultralow phase noise, flexible tuning capability, and high robustness. This is achieved through a quiet point（QP）-operated harmonic mode-locked fiber laser, which effectively mitigates optical amplifier noise and supermode competition, thus significantly diminishing the necessity for ultra-narrow electronic filters. Due to the liberated tuning ability, we present an oscillator that can be tuned from 2 GHz to 18 GHz, with phase noise as low as-140 d Bc/Hz at 10 k Hz under the QP operation. We then illustrate the practical application of the proposed oscillator in generating frequency-hopping signals with consistent spurious modes less than-85 d Bc,absolute phase noise below-135 d Bc/Hz at 10 k Hz, hopping resolution of 1.25 MHz, and fractional frequency stability below 6.1 × 10-12at 1 s averaging time when locked to a reference. The presented COEO structure emerges as a compelling solution for agile and low-noise microwave sources in advanced wireless communication and radar systems.

关键词： Bandpass filters Beam splitters Microwave photonics Mode-locked fiber lasers Optical amplifiers Phase noise

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Floquet spectrum and universal dynamics of a periodically driven two-atom system

引用

Chinese Physics B 2024年第2期33卷 436-443页

作者：谢文柱周正强李轩崔思淼孙明远 School of Science Beijing University of Posts and TelecommunicationsBeijing 100876China

We investigate the Floquet spectrum and excitation properties of a two-ultracold-atom system with periodically driven interaction in a three-dimensional harmonic *** interaction between the atoms is changed by varying the s-wave scattering length in two ways,the cosine and the square-wave *** is found that as the driving frequency increases,the Floquet spectrum exhibits two main features for both modulations,the accumulating and the spreading of the quasienergy levels,which further lead to different dynamical *** accumulation is associated with collective excitations and the persistent growth of the energy,while the spread indicates that the energy is bounded at all *** initial scattering length,the driving frequency and amplitude can all significantly change the Floquet spectrum as well as the ***,the corresponding relation between them is valid ***,we propose a mechanism for selectively exciting the system to one specific state by using the avoided crossing of two quasienergy levels,which could guide preparation of a desired state in experiments.

关键词： Floquet spectrum universal dynamics two-atom system avoided crossing

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

A credibility-aware swarm-federated deep learning framework in internet of vehicles

引用

Digital Communications and Networks 2024年第1期10卷 150-157页

作者： Zhe Wang Xinhang Li Tianhao Wu Chen Xu Lin Zhang School of Artificial Intelligence Beijing University of Posts and TelecommunicationsBeijing100876China

Although Federated Deep Learning(FDL)enables distributed machine learning in the Internet of Vehicles(IoV),it requires multiple clients to upload model parameters,thus still existing unavoidable communication overhead and data privacy *** recently proposed Swarm Learning(SL)provides a decentralized machine learning approach for unit edge computing and blockchain-based coordination.A Swarm-Federated Deep Learning framework in the IoV system(IoV-SFDL)that integrates SL into the FDL framework is proposed in this *** IoV-SFDL organizes vehicles to generate local SL models with adjacent vehicles based on the blockchain empowered SL,then aggregates the global FDL model among different SL groups with a credibility weights prediction *** experimental results show that compared with the baseline frameworks,the proposed IoV-SFDL framework reduces the overhead of client-to-server communication by 16.72%,while the model performance improves by about 5.02%for the same training iterations.

关键词： Swarm learning Federated deep learning Internet of vehicles Privacy Efficiency

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Improved Twin Support Vector Machine Algorithm and Applications in Classification Problems

引用

China Communications 2024年第5期21卷 261-279页

作者： Sun Yi Wang Zhouyang beijing university of posts and telecommunications School of Computer Science(National Pilot Software Engineering School) beijing university of posts and telecommunications International School

The distribution of data has a significant impact on the results of *** the distribution of one class is insignificant compared to the distribution of another class,data imbalance *** will result in rising outlier values and ***,the speed and performance of classification could be greatly *** the above problems,this paper starts with the motivation and mathematical representing of classification,puts forward a new classification method based on the relationship between different classification *** with the vector characteristics of the actual problem and the choice of matrix characteristics,we firstly analyze the orderly regression to introduce slack variables to solve the constraint problem of the lone *** we introduce the fuzzy factors to solve the problem of the gap between the isolated points on the basis of the support vector *** introduce the cost control to solve the problem of sample ***,based on the bi-boundary support vector machine,a twostep weight setting twin classifier is *** can help to identify multitasks with feature-selected patterns without the need for additional optimizers,which solves the problem of large-scale classification that can’t deal effectively with the very low category distribution gap.

关键词： fuzzy ordered regression(OR) relaxing variables twin support vector machine

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

On the Strong Approximation for a Simple Reentrant Line in Light Traffic Under First-buffer First-served Service Discipline

引用

Acta Mathematicae Applicatae Sinica 2024年第3期40卷 823-839页

作者： Kai-ming YANG Yong-jiang GUO School of Science Beijing University of Posts and TelecommunicationsBeijing 100876China

For a 2-station and 3-class reentrant line under first-buffer first-served(FBFS)service discipline in light traffic,we firstly construct the strong approximations for performance measures including the queue length,workload,busy time and idle time *** on the obtained strong approximations,we use a strong approximation method to find all the law of the iterated logarithms(LILs)for the above four performance measures,which are expressed as some functions of system parameters:means and variances of interarrival and service times,and characterize the fluctuations around their fluid approximations.

关键词： reentrant line queueing network FBFS service discipline strong approximation LIL Brownian motion

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Variability Analysis for a Two-station Queueing Network in Heavy Traffic with Arrival Processes Driven by Queues

引用

Acta Mathematicae Applicatae Sinica 2024年第2期40卷 445-466页

作者： Jian CAO Yong-jiang GUO Kai-ming YANG School of Science Beijing University of Posts and TelecommunicationsBeijing 100876China

The law of the iterated logarithm(LIL)for the performance measures of a two-station queueing network with arrivals modulated by independent queues is developed by a strong approximation *** convenience,two arrival processes modulated by queues comprise the external system,all others are belong to the internal *** is well known that the exogenous arrival has a great influence on the asymptotic variability of performance measures in *** the considered queueing network in heavy traffic,we get all the LILs for the queue length,workload,busy time,idle time and departure processes,and present them by some simple functions of the primitive *** LILs tell us some interesting insights,such as,the LILs of busy and idle times are zero and they reflect a small variability around their fluid approximations,the LIL of departure has nothing to do with the arrival process,both of the two phenomena well explain the service station’s situation of being busy all the *** external system shows us a distinguishing effect on the performance measures:an underloaded(overloaded,critically loaded)external system affects the internal system through its arrival(departure,arrival and departure together).In addition,we also get the strong approximation of the network as an auxiliary result.

关键词： two-station queueing network the law of the iterated logarithm(LIL) strong approximation fluid approximation

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Resource Allocation for Cognitive Network Slicing in PD-SCMA System Based on Two-Way Deep Reinforcement Learning

引用

China Communications 2024年第6期21卷 53-68页

作者： Zhang Zhenyu Zhang Yong Yuan Siyu Cheng Zhenjie School of Electronic Engineering Beijing University of Posts and TelecommunicationsBeijing 100876China beijing Key Laboratory of Work Safety Intelligent Monitoring(beijing university of posts and telecommunications) Beijing 100876China

In this paper,we propose the Two-way Deep Reinforcement Learning(DRL)-Based resource allocation algorithm,which solves the problem of resource allocation in the cognitive downlink network based on the underlay *** users(SUs)in the cognitive network are multiplexed by a new Power Domain Sparse Code Multiple Access(PD-SCMA)scheme,and the physical resources of the cognitive base station are virtualized into two types of slices:enhanced mobile broadband(eMBB)slice and ultrareliable low latency communication(URLLC)*** design the Double Deep Q Network(DDQN)network output the optimal codebook assignment scheme and simultaneously use the Deep Deterministic Policy Gradient(DDPG)network output the optimal power allocation *** objective is to jointly optimize the spectral efficiency of the system and the Quality of Service(QoS)of *** results show that the proposed algorithm outperforms the CNDDQN algorithm and modified JEERA algorithm in terms of spectral efficiency and QoS ***,compared with the Power Domain Non-orthogonal Multiple Access(PD-NOMA)slices and the Sparse Code Multiple Access(SCMA)slices,the PD-SCMA slices can dramatically enhance spectral efficiency and increase the number of accessible users.

关键词： cognitive radio deep reinforcement learning network slicing power-domain non-orthogonal multiple access resource allocation

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Dual-band dynamic reconfigurable intelligent surface for signal enhancement in millimeter wave

引用

Science China(Technological Sciences) 2024年第4期67卷 1133-1138页

作者： YAO HuiMing DU XinRun WANG HengBin GAO Song PANG YaChen ZHENG HaiRong GONG HuiWen LI Yi XU JianChun ZHANG JianHua BI Ke State Key Laboratory of Information Photonics and Optical Communications School of ScienceBeijing University of Posts and TelecommunicationsBeijing 100876China beijing Key Laboratory of Intelligent telecommunications Software and Multimedia Beijing University of Posts and TelecommunicationsBeijing 100876China State Key Laboratory of Networking and Switching Technology Beijing University of Posts and TelecommunicationsBeijing 100876China

Millimeter wave with large bandwidth,high transmission rate,and low delay is considered a reliable alternative to cope with the spectrum ***,the fast attenuation and narrow beam characteristics make it difficult to achieve long-distance or wide-range ***,a 1-bit dual-band reflective reconfigurable intelligent surface(RIS)for signal enhancement in millimeter wave with 16×16 elements is designed,fabricated,and *** from most existent RIS,dynamic programming is realized at two separate frequency bands by integrating the PIN diodes and field-programmable gate array(FPGA).Particularly,the beam deflection,dual-beam,and multi-beam are created based on the coding theory and convolution operation,proving the effectiveness of wavefront ***,the far-field patterns and signal power with different coding sequences are measured and *** is indicated that the received signal power is 6–7 dB stronger than that without coding,which shows good agreement with the desired *** proposed reconfigurable metasurface exhibits great potential in beam forming,making it a promising candidate for progressive wireless communication applications.

关键词： dual-band millimeter wave reconfigurable intelligent surface signal enhancement

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

FlowGANAnomaly: Flow-Based Anomaly Network Intrusion Detection with Adversarial Learning

引用

Chinese Journal of Electronics 2024年第1期33卷 58-71页

作者： Zeyi LI Pan WANG Zixuan WANG School of Computer Science Nanjing University of Posts and Telecommunications School of Modern posts Nanjing University of Posts and Telecommunications School of Internet of Things Nanjing University of Posts and Telecommunications

In recent years, low recall rates and high dependencies on data labelling have become the biggest obstacle to developing deep anomaly detection(DAD) techniques. Inspired by the success of generative adversarial networks(GANs) in detecting anomalies in computer vision and imaging, we propose an anomaly detection model called Flow GANAnomaly for detecting anomalous traffic in network intrusion detection systems(NIDS). Unlike traditional GAN-based approaches, which are composed of a flow encoder, a convolutional encoder-decoder-encoder, a flow decoder and a convolutional encoder, the architecture of this model consists of a generator(G) and a discriminator(D).Flow GANAnomaly maps the different types of traffic feature data from separate datasets to a uniform feature space,thus can capture the normality of network traffic data more accurately in an adversarial manner to mitigate the problem of the high dependence on data labeling. Moreover, instead of simply detecting the anomalies by the output of D, we proposed a new anomaly scoring method that integrates the deviation between the output of two Gs’ convolutional encoders with the output of D as weighted scores to improve the low recall rate of anomaly detection. We conducted several experiments comparing existing machine learning algorithms and existing deep learning methods(Auto Encoder and VAE) on four public datasets(NSL-KDD, CIC-IDS2017, CIC-DDo S2019, and UNSW-NB15). The evaluation results show that Flow GANAnomaly can significantly improve the performance of anomaly-based NIDS.

关键词： Deep learning Machine learning algorithms Computational modeling Network intrusion detection Imaging Telecommunication traffic Generative adversarial networks

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

在线全文

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：