检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Liu, Yingzhuo Zuo, Xingquan School of Computer Science Beijing University of Posts and Telecommunications Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education China

Multiple Line Bus Scheduling Problem (MLBSP) is vital to save operational cost of bus company and guarantee service quality for passengers. Existing approaches typically generate a bus scheduling scheme in an offline manner and then schedule buses according to the scheme. In practice, uncertain events such as traffic congestion occur frequently, which may make the pre-determined bus scheduling scheme infeasible. In this paper, MLBSP is modeled as a Markov Decision Process (MDP). A Reinforcement Learning-based Multi-line bus Scheduling Approach (RL-MSA) is proposed for bus scheduling at both the offline and online phases. At the offline phase, deadhead decision is integrated into bus selection decision for the first time to simplify the learning problem. At the online phase, deadhead decision is made through a time window mechanism based on the policy learned at the offline phase. We develop several new and useful state features including the features for control points, bus lines and buses. A bus priority screening mechanism is invented to construct bus-related features. Considering the interests of both the bus company and passengers, a reward function combining the final reward and the step-wise reward is devised. Experiments at the offline phase demonstrate that the number of buses used of RL-MSA is decreased compared with offline optimization approaches. At the online phase, RL-MSA can cover all departure times in a timetable (i.e., service quality) without increasing the number of buses used (i.e., operational cost). Copyright © 2024, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Multi-Agent Trust Evaluation Model based on Reinforcement Learning 8

Multi-Agent Trust Evaluation Model based on Reinforcement Le...

引用

8th International Conference on Dependable Systems and Their Applications, DSA 2021

作者： Haoran, Jia Yuyu, Yuan Qiang, Han Pengqian, Zhao Ting, Guo Ministry of Education Key Laboratory of Trustworthy Distributed Computing and Service Beijing University of Posts and Telecommunications China

ISBN: (纸本)9781665443913

The establishment of trust between agents is an important part of Multi-agent Reinforcement Learning (MARL). Recent studies on multi-agents have paid more attention to the game relationship between agents, ignoring trust is the basis of cooperation among agents. In this article, we propose a trust evaluation model in MARL to evaluate the trustworthiness of the target agent. The trust evaluation of the target agent is composed of direct trust and reputation. Direct trust is obtained from the historical interaction of agents. Reputation is to evaluate the impact of an agent's behavior by counterfactual action of a high-ranking agent. In addition, this paper also proposes a trust-based reward mechanism in MARL, which divides agent rewards into environmental rewards and trust rewards. Trust rewards are obtained through trust evaluation values, which drive the agents to reach a cooperative state and obtain higher long-term rewards. In this paper, we set up experiments in harvest game to verify that the trust evaluation model can effectively promote the cooperation of agents and avoid the problem of agents facing social dilemmas. © 2021 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

DACTI: a Generation-Based Data Augmentation Method for Cyber Threat Intelligence

DACTI: a Generation-Based Data Augmentation Method for Cyber...

引用

IEEE International Conference on Data Science in Cyberspace (DSC)

作者： Jingbing Chen Yali Gao Xiaoyong Li Key Lab of Trustworthy Distributed Computing and Service (BUPT) Ministry of Education School of Cyberspace Security Beijing University of Posts and Telecommunications Beijing China

Methods of data augmentation are extensively utilized to boost the efficacy of machine learning, given their superior capability in promoting generalization performance. In this paper, we address the issue of data scarcity in the field of cyber threat intelligence and propose a method for generating high-quality synthetic data by training a language model on linearized labeled phrases. Subsequently, we design a relational classification model to classify relational entities related to cyber threat intelligence in the created synthetic data. We conduct extensive experiments on datasets of different sizes, and the experimental results show that our proposed method outperforms the baseline methods, especially when the amount of gold data provided is insufficient, and the $F_{1}$ score of our model is improved by up to 1.94 points.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Novel Evolutionary Neural Network-based Approach for Online Bus Scheduling

A Novel Evolutionary Neural Network-based Approach for Onlin...

引用

IEEE International Conference on Networking, Sensing and Control

作者： Wenyang Tian Feiyang Wang Xingquan Zuo School of Computer Science Beijing University of Posts and Telecommunications Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education Beijing China

ISBN: (数字)9798350365221

ISBN: (纸本)9798350365238

Bus scheduling is vital to ensure service quality and save operational costs. Current studies typically consider the problem as an optimization problem and solve it using exact or heuristic approaches. In this paper, we consider the problem as a sequential decision-making problem and propose an Evolutionary Neural Network-based Approach (ENNA) for it. A neural network is used to make a decision at each departure time in the bus timetable. It selects which bus should depart at that specific time. By this means, a complete bus scheduling scheme is generated by a series of decisions made by the neural network. The structure and weights of the neural network are optimized by a genetic algorithm. To meet the real-world operational needs of bus duty types and improve bus utilization, we design a bus duty type transition rule that dynamically decides the duty type of buses. Experimental results demonstrate that the proposed ENNA outperforms manual scheduling schemes and can generate bus scheduling schemes with fewer buses.

关键词： Network topology Decision making Artificial neural networks Transforms Dynamic scheduling Real-time systems Topology Sensors Optimization Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Graph Neural Network Based User Group Comments Interaction Behavior Prediction on Microblog Websites 6

Graph Neural Network Based User Group Comments Interaction B...

引用

6th IEEE International Conference on Data Science in Cyberspace, DSC 2021

作者： Chen, Yisen Wu, Xu Xie, Xiaqing Wu, Jingchen Qiu, Lirong Sun, Lijuan Beijing University Of Posts And Telecommunications Ministry Of Education Key Laboratory Of Trustworthy Distributed Computing And Service Beijing China

ISBN: (纸本)9781665418157

The interactive behavior of comments within the user group on microblog website is bidirectional and dynamic, reflecting the level of familiarity among users. Predicting the future comment interaction behavior within the user group is of great significance for commercial recommendation and crime fighting. Related research groups regard it as a temporal link prediction problem, assume the evolution is smooth or there is no evolution at all. Meanwhile, the feature in those studies is too unitary, leading to low link prediction performance. In this paper, we propose a method called User Group Comment Interaction Behaviour Prediction (UGCIBP), which combines structure extraction layer and evolution extraction layer to do dynamic graph representation learning to acheive the modeling of user group historical comment interaction graphs, and build the features of communication weight, interest similarity, and common degree of activity between users on microblog website. Then, we combine dynamic graph representation learning and the constructed features to do link prediction, finally, the purpose of predicting comment interaction is acheived. Experiments are conducted on the public datasets Enron and Twitter datasets to evaluate the method. The results show that the AUC scores of this method are 4.74% and 9.95% higher than existing methods, respectively, which proves the effectiveness of the UGCIBP method. © 2021 IEEE.

关键词： Websites

来源：评论

学校读者我要写书评

暂无评论

Video Emotional Classification Based on Deep Reinforcement Learning

Video Emotional Classification Based on Deep Reinforcement L...

引用

Communications Technology and Computer Science (ACCTCS), Asia-Pacific Conference on

作者： Tingting Yuan Yuyu Yuan Key Laboratory of Trustworthy Distributed Computing and Service (BUPT) Ministry of Education Beijing University of Posts and Telecommunications Beijing China

We perform video emotion classification based on deep reinforcement learning, simulate the process of a human watching video and getting the emotional expression, and find a way to combine reinforcement learning algorithms with video emotion analysis characters. The Conv-LSTM network is built as an agent in the critical model, and the LSTM network is trained in combination with the A3C algorithm in reinforcement learning, which makes the model have emotion recognition ability and verifies the effectiveness on single and compound emotion datasets.

关键词：

来源：评论

学校读者我要写书评

暂无评论

OptimizeLog: Log Anomaly Detection and Localization based on Optimized Log Parsing in distributed Systems

OptimizeLog: Log Anomaly Detection and Localization based on...

引用

International Conference on Computer and Communications (ICCC)

作者： Yanjie Sun Yali Gao Xiaoyong Li Key Laboratory of Trustworthy Distributed Computing and Service (BUPT) Ministry of Education Beijing University of Posts and Telecommunications Beijing China

Modern distributed systems generate interleaved logs when performing parallel operations, and these logs become an important basis for anomaly detection and localization. To achieve more robust and accurate log anomaly detection and localization, we propose a comprehensive model, namely OptimizeLog. We built a dictionary- and length-based log parser that works stably in most log systems without parameter tuning. We significantly improve anomaly detection accuracy by introducing ALBERT-based semantic embeddings, count embeddings, and time embeddings, combined with an attentionbased Bi-GRU model. By constructing component instance forest and tree-based depth-first traversal algorithm, abnormal location in distributed systems is realized. Experimental results show that our log parsing method is 4% more accurate than other log parsers. Compared with other advanced methods such as DeepLog, LogAnomaly, and LogRobust, OptimizeLog improves the accuracy of anomaly detection by 5%, while enabling instancelevel anomaly localization on real datasets.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Clean-label Poisoning Attack against Fake News Detection Models

Clean-label Poisoning Attack against Fake News Detection Mod...

引用

IEEE International Conference on Big Data

作者： Jiayi Liang Xi Zhang Yuming Shang Sanchuan Guo Chaozhuo Li Key Laboratory of Trustworthy Distributed Computing and Service (MOE) Beijing University of Posts and Telecommunications Beihang University Beijing China

Researching data poisoning attacks against fake news detection models is crucial for bolstering their robustness and curbing the dissemination of fake news. Existing textual data poisoning attacks necessitate control over both the content and labels of news samples, making them impractical for real attack scenarios. In this paper, we propose COMCP, a novel clean-label poisoning attack model aimed at fake news detection models. Diverging from existing methods, COMCP ensures the poison samples are accurately labeled, while crafting stealthy poison comments without modifying the headlines or content, thereby enhancing the feasibility of the attack. Furthermore, COMCP generates poison comments by appending stealthy characters to ensure the stealthiness of the attack. Comprehensive experimental evaluations on three benchmark datasets illustrate that our proposal outperforms SOTA baselines in terms of attack success rate and text quality, while maintaining the accuracy of detecting clean samples.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Novel Low-Complexity Angle-Assisted Delay and Doppler Estimation for MIMO-OTFS

A Novel Low-Complexity Angle-Assisted Delay and Doppler Esti...

引用

International Conference on Communication Technology (ICCT)

作者： Weixin Deng Hui Gao Ruohan Cao Xinzhou Cheng Key Laboratory of Trustworthy Distributed Computing and Service MOE Beijing University of Posts and Telecommunications Beijing China China Unicom Research Institute Beijing China

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

Multiple-input multiple-output orthogonal time frequency space (MIMO-OTFS) is considered one of the potential technologies for high-speed mobile scenarios. However, in such systems, fractional delay and Doppler estimation is a major challenge due to the limitation of delay-Doppler (DD) grid resolution. In this paper, we propose a low-complexity angle-assisted delay and Doppler estimation (AADDE) algorithm for MIMO-OTFS systems. AADDE reduces computational complexity by estimating delay and Doppler parameters separately, and the angle information is used to improve estimation performance. In addition, we propose a two-dimensional (2D) grid segmentation scheme to estimate fractional DD with AADDE. Specifically, an iterative approach is developed to identify the true fractional DD index using a variable search step size, achieving lower complexity than existing off-grid search methods. Complexity analysis and simulation results demonstrate the effectiveness of the proposed AADDE algorithm, which also approaches the Cramér-Rao lower Bound.

关键词： Time-frequency analysis Lower bound Simulation Two-dimensional displays Search methods Estimation MIMO Delays Doppler effect Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Insider Threat Detection Based on User Historical Behavior and Attention Mechanism 6

Insider Threat Detection Based on User Historical Behavior a...

引用

6th IEEE International Conference on Data Science in Cyberspace, DSC 2021

作者： He, Weiyu Wu, Xu Wu, Jingchen Xie, Xiaqing Qiu, Lirong Sun, Lijuan Beijing University Of Posts And Telecommunications Ministry Of Education Key Laboratory Of Trustworthy Distributed Computing And Service Beijing China

ISBN: (纸本)9781665418157

Insider threat makes enterprises or organizations suffer from the loss of property and the negative influence of reputation. User behavior analysis is the mainstream method of insider threat detection, but due to the lack of fine-grained detection and the inability to effectively capture the behavior patterns of individual users, the accuracy and precision of detection are insufficient. To solve this problem, this paper designs an insider threat detection method based on user historical behavior and attention mechanism, including using Long Short Term Memory (LSTM) to extract user behavior sequence information, using Attention-based on user history behavior (ABUHB) learns the differences between different user behaviors, uses Bidirectional-LSTM (Bi-LSTM) to learn the evolution of different user behavior patterns, and finally realizes fine-grained user abnormal behavior detection. To evaluate the effectiveness of this method, experiments are conducted on the CMU-CERT Insider Threat Dataset. The experimental results show that the effectiveness of this method is 3.1% to 6.3% higher than that of other comparative model methods, and it can detect insider threats in different user behaviors with fine granularity. © 2021 IEEE.

关键词： Anomaly detection

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：