检索结果-内蒙古大学图书馆

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Lin, Jiayi Zhang, Chenyang Tong, Haibo Zhang, Dongyu Hong, Qingqing Hou, Bingxuan Wang, Junli Key Laboratory of Embedded System and Service Computing Tongji University Ministry of Education Shanghai201804 China Collaborative Innovation Center for Financial Network Security Tongji University Shanghai201804 China

ISBN: (纸本)9798891761681

Multi-Span Question Answering (MSQA) requires models to extract one or multiple answer spans from a given context to answer a question. Prior work mainly focuses on designing specific methods or applying heuristic strategies to encourage models to predict more correct predictions. However, these models are trained on gold answers and fail to consider the incorrect predictions. Through a statistical analysis, we observe that models with stronger abilities do not predict less incorrect predictions compared with other models. In this work, we propose Answering-Classifying-Correcting (ACC) framework, which employs a post-processing strategy to handle incorrect predictions. Specifically, the ACC framework first introduces a classifier to classify the predictions into three types and exclude "wrong predictions", then introduces a corrector to modify "partially correct predictions". Experiments on several MSQA datasets show that ACC framework significantly improves the Exact Match (EM) scores, and further analysis demostrates that ACC framework efficiently reduces the number of incorrect predictions, improving the quality of predictions. © 2024 Association for Computational Linguistics.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

A remote debugging system for OpenMP parallel program

A remote debugging system for OpenMP parallel program

引用

2011 International Conference on Transportation, Mechanical, and Electrical Engineering, TMEE 2011

作者： Chai, Hua Shi, Qinqin Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai 200092 China Department of Computer Science and Information Engineering Shanghai Institute of Technology Shanghai 201418 China

ISBN: (纸本)9781457717017

A remote debugging system for OpenMP parallel program is presented in this paper. The system consists of two parts, namely, an integrated debugging environment running on the clent-side and a background daemon running on the server-side. Information exchange between the two sides is accomplished through the socket-based network communication technology. The remote debugging function is realized by the automatic instrumentation technology at code level based on syntax tree. Instance tests show that the system can provide remote users with correctness checking, performance analysis and other debugging functions for OpenMP parallel programs. The user interface is simple and very easy to use. © 2011 IEEE.

关键词： User interfaces

来源：评论

学校读者我要写书评

暂无评论

An evolutionary semi-supervised subtractive clustering method by seeding

An evolutionary semi-supervised subtractive clustering metho...

引用

2012 International Conference on Information Technology and Software Engineering, ITSE 2012

作者： Gu, Lei Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai 200092 China School of Computer Science and Technology Nanjing University of Posts and Telecommunication Nanjing 210046 China

ISBN: (纸本)9783642345210

A semi-supervised subtractive clustering has been proposed recently. However, it performance depends greatly on the choice of the parameters of the mountain function and only proper parameters enable the clustering method to produce a better effect. In this paper, an evolutionary semi-supervised subtractive clustering method by seeding is presented. The novel approach uses the genetic algorithms for tuning the relative parameters of the subtractive clustering by minimizing the Davies Bouldin clustering validity index regarded as the fitness function. To investigate the effectiveness of our approach, several experiments are done three real datasets. Experimental results show that our proposed method can improve the clustering performance significantly compared to other two traditional semi-supervised clustering algorithms. © 2013 Springer-Verlag.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Exploiting robustness in Petri net-based closed-loop systems to accommodate control specification changes

Exploiting robustness in Petri net-based closed-loop systems...

引用

2013 IEEE International Conference on Automation Science and Engineering, CASE 2013

作者： Li, Jun Zhou, Mengchu Dai, Xianzhong Key Laboratory of Measurement and Control of Complex Systems of Engineering China Ministry of Education Automation School of Southeast University Nanjing 210096 China Key Laboratory of Embedded System and Service Computing Ministry of Education Tongji University Shanghai 201804 China

ISBN: (纸本)9781479915156

Supervisory control reconfiguration has been paid more attention due to its capacity of handling the uncertainties including resource failures, capability degradation, and task changes in discrete event systems. However, the robustness of closed-loop systems in response to some specification changes was not addressed in the prior studies. This paper presents a reconfiguration method based on Petri nets and integer programming for supervisory control of resource allocation systems with changeable resource allocation relationships (control specifications). First, we propose a reconfiguration mechanism. It updates the P-invariant-based supervisor and evolves the state of the closed-loop system. The latter adapts to the certain changes in specifications by the self-regulation of the closed-loop system's state. It utilizes fully the robustness in a closed-loop system and avoids unnecessary modification of a supervisor. Then, a reconfiguration model with robustness analysis is proposed based on integer programming. By using it, a firing sequence of transitions as a serial of reconfiguration actions can be found. The execution of the sequence implements the state evolution of a closed-loop system. There is no need for making additional efforts to analyze state reachability of the system, since the firability condition of transitions is incorporated into the model. Finally, the method is illustrated through an example of a hospital emergency service system. © 2013 IEEE.

关键词： Closed loop systems

来源：评论

学校读者我要写书评

暂无评论

ADHD Discrimination Based on Social Network

ADHD Discrimination Based on Social Network

引用

International Conference on Cloud computing and Big Data (CCBD)

作者： Xiaojiao Guo Lianghua He The key Laboratory of Embedded System and Service Computing Tongji University Shanghai China

ISBN: (纸本)9781479966226

Attention Deficit Hyperactivity Disorder (ADHD) is one of the common diseases of brain and has brought the growth of teenagers and even the adult indelible damage. It is very different to classify the ADHD symptoms and normal by the existing research. In this paper, the contributions are as two aspects: one is that the attributes of brain network of the resting state fMRI data have been calculated to discriminate three categories ADHD from the controls. And the average accuracies of various categories are 42.49% and 63.46% on the ADHD-200 datasets of NYU and KKI respectively, which is better than the average best imaging-based diagnostic performance of 35.19% and 61.90% achieved in the ADHD-200 global competition. The other one is that we put forward a new method named G-algorithm to construct the whole brain network, which based on certain rules. The same or even better classification results have been achieved by this method, which also verifies its feasibility and effectiveness.

关键词： Social network services Accuracy Support vector machines Correlation Synchronization Diseases Organizations

来源：评论

学校读者我要写书评

暂无评论

Entropy-based Traffic Congestion Propagation Pattern Mining with GPS Data

Entropy-based Traffic Congestion Propagation Pattern Mining ...

引用

IEEE International Conference on Big Data Analysis (ICBDA)

作者： Xinyuan Sui Yaying Zhang Key Laboratory of Embedded System and Service Computing Tongji University Shanghai China

Traffic congestion in urban areas has become a critical issue in the world. As the limits of road topology and the influence of residents' commuting periodic patterns, the same part of the road network usually gets congested, and there are similar propagation paths when congestion occurs in an area. This paper aims to model the congestion propagation phenomenon and further discover the congestion propagation paths that frequently appeared. To this end, we propose an algorithm which analyzes the correlations between congestions on different road segments and constructs a congestion propagation graph to reflect the dynamic process of congestions. The traffic flow entropy is novelty considered to identify the potential essential road segment in traffic congestions. We adopt the Jaccard correlation to estimate the parameters employed. The experiments with the road network in Beijing, China and real-world GPS trajectory data indicate that our approach can reveal congestion propagation patterns and the dynamic process of congestion propagation.

关键词： Correlation Network topology Roads Heuristic algorithms Urban areas Trajectory Topology

来源：评论

学校读者我要写书评

暂无评论

Local-directed network coding in vehicular ad-hoc networks

Local-directed network coding in vehicular ad-hoc networks

引用

IEEE International Symposium on Intelligent Signal Processing and Communication systems (ISPACS)

作者： Haojie Luo Dongdong Zhang Chen Ye The Key Laboratory of Embedded System and Service Computing Tongji University Shanghai China

In this paper, we proposed a novel network coding method called local-directed network coding to improve the throughput of the vehicular ad-hoc network. Different from other vehicular network coding methods, we consider the direction of packet transmission. In detail, the intermediate node selects one main packet and some sub-packets from forwards and backwards sending queues, and then encodes these packets into one packet to broadcast. Simulation results show that our approach can efficiently improve the throughput of application layer in vehicular networks.

关键词： Encoding Buffer storage Accidents Vehicles Streaming media Artificial neural networks Performance evaluation

来源：评论

学校读者我要写书评

暂无评论

Counting Pedestrian in Crowded Subway Scene

Counting Pedestrian in Crowded Subway Scene

引用

Congress on Image and Signal Processing, CISP

作者： Keju Zu Fuqiang Liu Zhipeng Li Key Laboratory of Embedded System and Service Computing University of Tongji Shanghai China

When the high occlusion occurs in crowded scene, face detection is a better substitute for detecting pedestrian. In this paper, we present a novel crowd analysis method based on discriminative descriptor of faces and support vector machine (SVM) ensemble. Through manipulating the input features in the same sample set, the different input features of faces are extracted to train two SVM classifiers. The classification scores of two generated classifiers are combined adaptively to make a collective decision. The first SVM, as the principal classifier gives out most of face hypotheses, while the second SVM serves as secondary one to rejecting the false positive. We present experiment to test the proposed method in crowded subway video, and the result shows that the SVM ensemble outperforms the single SVM in counting the pedestrian.

关键词： Layout Support vector machines Support vector machine classification Face detection Histograms Surveillance Humans Lighting Cameras Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Image Retrieval Based on Manifold Learning and Incorporate Clustering

Image Retrieval Based on Manifold Learning and Incorporate C...

引用

GCIS, WRI Global Congress on Intelligent systems

作者： Jianzhu Cui Fuqiang Liu Zhipeng Li Jing Li The Key Laboratory of Embedded System and Service Computing University of Tongji Shanghai China

This paper presents a novel unsupervised learning framework named image retrieval based on manifold learning and incorporate clustering. The dimensionality of image descriptors used in image retrieval applications is quite high. Given a query image, our algorithm first makes use of manifold learning (LPP) for dimensionality reduction and manifold ranking algorithm to explore the relationship among all the data points in the feature space, and then measures relevance between the query and all the images in the database accordingly. Then we use the similarities among target images for improving the performance of the image retrieval systems by cluster-based retrieval of images by unsupervised learning. Our algorithm retrieves image clusters as retrieval results by applying K-means clustering algorithm to a collection of images collected by manifold ranking algorithm. Experimental results on a general-purpose image database show that our algorithm attains a significant improvement over existing systems.

关键词： Image retrieval Clustering algorithms Feedback Unsupervised learning Support vector machines Support vector machine classification Image databases Content based retrieval Intelligent systems Laboratories

来源：评论

学校读者我要写书评

暂无评论

Reliable web services selection based on finite state machine model

引用

Information Technology Journal 2011年第9期10卷 1662-1672页

作者： Shen, Hongjie Ding, Zhijun Chen, Hongzhong Key Laboratory of Embedded System and Service Computing Tongji University Ministry of Education Shanghai 200092 China

Dynamically selecting suitable Web services (WSs) is crucial to users in Web services Composition (WSC). Generally, most works regard a Web service (WS) as the basic unit and compose the composite WS (CWS) end to end. However, a WS may comprise multiple operations that are invoked in sequence and the selection of WSs according to the model still is a NP problem. In this study, the WSs consist of some operations and then WSs selection problem is formalized as a Finite State Machine (FSM) process. This study proposed an algorithm to create the WSC Tree (WSCT), in which each path from root to leaf node is a feasible WSC execution path. Then, a heuristic algorithm is proposed to realize the selection of WSs based on the WSCT. The advantage of heuristic algorithm is that it can address the selection problem without traversing the whole WSCT but it still spends much time in the construction of the WSCT. So, in order to further decrease the time of selection, a backtracking algorithm is presented to select a feasible execution path without generating WSCT. At last, experiments show that heuristic algorithm is more effective than exhaustive method and the backtracking algorithm spend the less time than heuristic algorithm, the reliability of WSC execution path selected by backtracking algorithm can satisfy the requirement of users. So, this study proposed an algorithm to select a feasible WSC execution path within a reasonable time. © 2011 Asian Network for Scientific Information.

关键词： Web services

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：