检索结果-内蒙古大学图书馆

Digital Society and Intelligent Systems (DSInS), International Conference on

作者： Zhilin Liu Hengwei Zhang Pengyu Sun Yan Mi Xiaoning Zhang Jindong Wang State Key Laboratory of Mathematical Engineering and Advanced Computing Key Laboratory of Information Security of Henan Provincial Zhengzhou China

ISBN: (纸本)9781665406314

At present, APT attack detection mainly focuses on the specific realization of detection technology, and there is little research on the timing strategy for active detection of APT attack. This paper focuses on the problem of APT dynamic detection with timing decision intelligence. We use the FlipIt game model and reinforcement learning method of Q-learning algorithm, consider the APT attack events as an exponentially distributed random process, and the defender using the attacker’s last move time (LM-defender) to learn his optimal timing strategy using Q-Learning algorithm. The defender's reinforcement learning mechanism for the timing of the APT attack action and outputs its timing selection strategy. Simulation experiments show that compared with the existing greedy algorithm, the timing strategy based on the Q-Learning algorithm does not require the attacker's prior information under the premise, it can achieve a higher average benefit than the greedy strategy. This result provides a theoretical reference for the design of timing strategy to defend APT attack intelligently and efficient.

关键词： Greedy algorithms Adaptation models state feedback Q-learning Adaptive systems Heuristic algorithms Games

来源：评论

学校读者我要写书评

暂无评论

Network Attack Intention Recognition Based on Signaling Game Model and Netlogo Simulation

Network Attack Intention Recognition Based on Signaling Game...

引用

Digital Society and Intelligent Systems (DSInS), International Conference on

作者： Xiaoning Zhang Hengwei Zhang Chenwei Li Pengyu Sun Zhilin Liu Jindong Wang State Key Laboratory of Mathematical Engineering and Advanced Computing Key Laboratory of Information Security of Henan Provincial Zhengzhou China

ISBN: (纸本)9781665406314

With the development of attack technology, identifying the real intention of network attack and avoiding further deterioration of security situation has become a hot research topic in recent years. This paper proposes an attack intention recognition method based on signaling game model. Firstly, we generate attack intention hypotheses, and then calculate the probability of each attack intention by solving game equilibrium. Experiments show that the method can effectively improve the accuracy of attack intention recognition from the perspective of the revenue.

关键词： Games Probability Cognition Complexity theory Security Intelligent systems Game theory

来源：评论

学校读者我要写书评

暂无评论

Binary semantic Similarity Comparison based on software gene

Binary semantic Similarity Comparison based on software gene

引用

作者： Wenjie Sun Yihang Chen Zheng Shan Fudong liu Qiuna Wang State Key Laboratory of Mathematical Engineering and Advanced Computing Materials branch of State Grid Henan Electric Power Company

The similarity of binary code is widely used in code copyright protection, vulnerability mining, malicious code analysis and etc. In this paper, we proposed a method for measuring/evaluating the similarity of two binary files based on software genes. Some of Natural language processing methods were introduced into program semantics analysis, including word2vec and doc2vec models to generate assembly instruction embeddings and gene semantic embeddings. Then the longest common subsequence method was applied to evaluate the software similarity. Experiments show that our method can effectively evaluate the similarity of binary files.

关键词： Binary semantic Similarity Comparison based on software gene

来源：评论

学校读者我要写书评

暂无评论

LLMEmbed: Rethinking Lightweight LLM's Genuine Function in Text Classification

arXiv

引用

arXiv 2024年

作者： Liu, Chun Zhang, Hongguang Zhao, Kainan Ju, Xinghai Yang, Lin Systems Engineering Institute AMS Beijing China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

With the booming of Large Language Models (LLMs), prompt-learning has become a promising method mainly researched in various research areas. Recently, many attempts based on prompt-learning have been made to improve the performance of text classification. However, most of these methods are based on heuristic Chain-of-Thought (CoT), and tend to be more complex but less efficient. In this paper, we rethink the LLM-based text classification methodology, propose a simple and effective transfer learning strategy, namely LLMEmbed, to address this classical but challenging task. To illustrate, we first study how to properly extract and fuse the text embeddings via various lightweight LLMs at different network depths to improve their robustness and discrimination, then adapt such embeddings to train the classifier. We perform extensive experiments on publicly available datasets, and the results show that LLMEmbed achieves strong performance while enjoys low training overhead using lightweight LLM backbones compared to recent methods based on larger LLMs, i.e. GPT-3, and sophisticated prompt-based strategies. Our LLMEmbed achieves adequate accuracy on publicly available benchmarks without any fine-tuning while merely use 4% model parameters, 1.8% electricity consumption and 1.5% runtime compared to its counterparts. Code is available at: https://***/ChunLiu-cs/LLMEmbed-ACL2024. © 2024, CC BY-NC-SA.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Data Dependence Constraints on SIMD Extensions

Data Dependence Constraints on SIMD Extensions

引用

2014 International Conference on Computer,Network Security and Communication engineering(CNSCE 2014)

作者： Yong-sheng HOU Rong-cai ZHAO Shi-jie YAN Xuan LI State Key Laboratory of Mathematical Engineering and Advanced Computing School of Automation and Electrical Engineering Beijing University of Science and Technology

Exploiting fine-grained data-level parallelism via SIMD(Single Instruction Multiple Data) extension instructions is a universal approach in recent decades. A parallelizing compiler with SIMD vectorization is a widespread admitted tool to utilize SIMD extensions, but it cannot exploit the inherent parallelisms of a serial program completely. In this paper, we analyze the data dependence constraints on SIMD extensions, and propose a Tarjan algorithm based SIMD vectorization detection prototype. This proposed detection prototype can recognize more inherent parallelisms of applications in practice. The experimental results show that the ability of the proposed prototype is prior to that of the ICC v10.0 compiler with respect to complicated data dependence constraints on SIMD extensions.

关键词： SIMD dependence analysis Tarjan

来源：评论

学校读者我要写书评

暂无评论

DSSA: Dual-Side Sparse Systolic Array Architecture for Accelerating Convolutional Neural Network Training 22

DSSA: Dual-Side Sparse Systolic Array Architecture for Accel...

引用

Proceedings of the 51st International Conference on Parallel Processing

作者： Zhengbo Chen Qi Yu Fang Zheng Feng Guo Zuoning Chen Information Engineering University China State Key Laboratory of Mathematical Engineering and Advanced Computing China Chinese Academy of Engineering China

ISBN: (纸本)9781450397339

Ever-growing CNN size incurs a significant amount of redundancy in model parameters, which in turn, puts considerable burden on hardware. Unstructured pruning is widely used to reduce model sparsity. While, the irregularity introduced by unstructured pruning makes it difficult to accelerate sparse CNNs on systolic array. To address this issue, a variety of accelerators have been proposed. SIGMA, the state-of-the-art sparse GEMM accelerator, achieves significant speedup over systolic array. However, SIGMA suffers from two disadvantages: 1) it only supports one-side sparsity, leaving potential for further performance gains; 2) SIGMA improves utilization of large-sized systolic arrays at the cost of extra overhead. In this paper, we propose DSSA, a dual-side sparse systolic array, to accelerate CNN training. DSSA bases its designs on a small-sized systolic array, which naturally achieves higher cell utilization without additional overhead. To facilitate dual-side sparsity processing, DSSA utilizes a cross-cycle reduction module to accumulate partial sum that belongs to the same column but being processed in different cycles. A comprehensive design space exploration is performed to seek the local optimal configurations for DSSA. We implement the logic design of DSSA using Verilog in RTL and evaluate its performance using a C++-based cycle-accurate performance simulator we built. Experimental results show that DSSA delivers, on average, a speedup of 2.13x and 13.81x over SIGMA and a basic systolic array with the same number of cells. Compared to SIGMA, DSSA incurs 16.59% area overhead and 25.49% power overhead when sparse filter is excluded, as SIGMA did.

关键词： Convolutional Neural Network Hardware Accelerator Sparsity Processing Systolic Array Unstructured Pruning

来源：评论

学校读者我要写书评

暂无评论

A practical and aggressive loop fission technique 18th

A practical and aggressive loop fission technique

引用

18th International Conference on Algorithms and Architectures for Parallel Processing - ICA3PP 2018 Collocated Workshops: ICA3PP 2018 Workshop on Intelligent Algorithms for Large-scale Complex Optimization Problems, IALCOP 2018, ICA3PP 2018 Workshop on Security and Privacy in Data Processing, SPDP 2018

作者： Zhao, Bo Li, Yingying Han, Lin Zhao, Jie Gao, Wei Zhao, Rongcai Yahyapour, Ramin Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen Göttingen37077 Germany State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou450001 China French Institute for Research in Computer Science and Automation Rocquencourt France

ISBN: (纸本)9783030052331

Loop fission is an effective loop optimization for exploiting fine-grained parallelism. Currently, loop fission is widely used in existing parallelizing compilers. To fully exploit the optimization, we proposed and implemented a practical and aggressive loop fission technique. First, we present an aggressive dependence graph pruning method to eliminate pseudo dependences caused by the conservativeness of compilers. Second, we introduce a topological sort based loop fission algorithm to distribute loops correctly. Finally, to enhance the performance of the generated programs which have potential of loop fission, we propose an advanced loop fission strategy. We evaluate these techniques and algorithms in the experimental section. © Springer Nature Switzerland AG 2018.

关键词： Program compilers

来源：评论

学校读者我要写书评

暂无评论

SARF: A Spark Based Recommendation Framework in the Cloud

SARF: A Spark Based Recommendation Framework in the Cloud

引用

International Conference on advanced Cloud and Big Data (CBD)

作者： Jieyao Zhou Yifei Lu Ka Bian Lei Qian School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China State Key Lab of Mathematical Engineering and Advanced Computing Wuxi China

With the development of the IT industry, the information resources and services that people come into contact with have been growing. The sharing and interaction of these two types of information has led to further growth in data resulting in the advent of the era of big data. However, due to the complexity of information, such as low utilization, the confusion, and disorder, it will be a great challenge for service providers to offer customers services with high quality and accurate quality. The recommendation system is a solution adopted by e-marketers for such problems. This paper proposes a SPARK-based recommendation algorithm called SARF. The SARF first reads the original data from the HDFS and uses Spark to yield the datasets. Then, by using these datasets, the optimal recommended model is generated through iterative process. Finally, we gain the results which are recommended to the user by the optimal model. The final experimental results show that our SARF framework can effectively reduce the recommendation error and solve the problem of cold start.

关键词： Collaboration Filtering Sparks Big Data Matrix decomposition Training Filtering algorithms

来源：评论

学校读者我要写书评

暂无评论

The interactive mechanism of static and dynamic analysis in the reverse analysis of embedded software

International Journal of Multimedia and Ubiquitous Engineeri...

引用

International Journal of Multimedia and Ubiquitous engineering 2016年第10期11卷 33-44页

作者： Tie-Ming, Liu Lie-Hui, Jinag Jing-Si, Zhu Gang, Meng The State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou450000 China ZhengZhou University of Science and Technology Zhengzhou450064 China

Because the software reverse analysis method which combined the dynamic and static analyses lacks normative interactive mode, the work of the software reverse analysis is inefficient, and its reusability is poor. Based on dynamic and static analysis process of the embedded software, three kinds of interactive mechanism are proposed, including Static To Dynamic (STD), Dynamic To Static (DTS), Static and Dynamic simultaneous (SDM), and has also presented the method of realizing these three interaction mechanisms in detail. The test results show that interactive mechanisms of STD, DTS and SDM are suitable for correction of abnormal nodes in the results of static analysis, optimization of dynamic information extraction, identification of hidden codes and so on. It can greatly improve work efficiency of the embedded software reverse analysis. © 2016 SERSC.

关键词： Static analysis

来源：评论

学校读者我要写书评

暂无评论

DCTCP with Front Marking: Decouple ECN Control Loop From The Queuing Delay

DCTCP with Front Marking: Decouple ECN Control Loop From The...

引用

IEEE International Conference on High Performance computing and Communications (HPCC)

作者： Xu Ma Yifei Lu Zhengzhi Xu School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China State Key Lab of Mathematical Engineering and Advanced Computing Wuxi China

Datacenter transports should provide low flow completion times (FCT) to achieve desired application performance for short flows, and it should also ensure largely sustained throughput for long flows. DCTCP is designed for datacenter to achieve those goals. As an end-to-end congestion algorithm, DCTCP marks packets, those exceed the threshold utilizing Explicit Congestion Notification (ECN) and feedbacks congestion extent to the sender for accurate congestion control. However, DCTCP marks the packet at the tail of the queue, which causes the ECN signal to experience hundreds of microseconds queue delay. In this paper, an enhanced DCTCP called DCTCP-FM (DCTCP with front marking) is designed based on the notion of ECN. DCTCP-FM ensures ECN marking on the front of queue. Thus, the mark always conveys the up-to-date information about the state of congestion at the time of packet departure, eliminating packets queuing delay. As a result, the sender can cut down its congestion window timely to alleviate congestion. Our preliminary simulation results show that compared to DCTCP, DCTCP-FM can: 1)put off the throughput collapse caused by Incast traffic; 2)reduce the buffer size requirement at the switch and avoid packet losses when the buffer is limited; 3)suppress queue oscillation to reduce end-to-end latency; 4) reduce 99th percentile FCTs of short flows in a real datacenter simulation.

关键词： Data centers Simulation High performance computing Conferences Packet loss Switches Jitter

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：