检索结果-内蒙古大学图书馆

arXiv 2023年

作者： He, Xingwei Zhang, Qianru Jin, A-Long Ma, Jun Yuan, Yuan Yiu, Siu Ming The University of Hong Kong Hong Kong School of Computer Science and Engineering Beihang University Beijing China State Key Laboratory of Software Development Environment China Zhongguancun Laboratory China

Factual error correction (FEC) aims to revise factual errors in false claims with minimal editing, making them faithful to the provided evidence. This task is crucial for alleviating the hallucination problem encountered by large language models. Given the lack of paired data (i.e., false claims and their corresponding correct claims), existing methods typically adopt the ‘mask-then-correct’ paradigm. This paradigm relies solely on unpaired false claims and correct claims, thus being referred to as distantly supervised methods. These methods require a masker to explicitly identify factual errors within false claims before revising with a corrector. However, the absence of paired data to train the masker makes accurately pinpointing factual errors within claims challenging. To mitigate this, we propose to improve FEC by Learning to Inject Factual Errors (LIFE), a three-step distantly supervised method: ‘mask-corrupt-correct’. Specifically, we first train a corruptor using the ‘mask-then-corrupt’ procedure, allowing it to deliberately introduce factual errors into correct text. The corruptor is then applied to correct claims, generating a substantial amount of paired data. After that, we filter out low-quality data, and use the remaining data to train a corrector. Notably, our corrector does not require a masker, thus circumventing the bottleneck associated with explicit factual error identification. Our experiments on a public dataset verify the effectiveness of LIFE in two key aspects: Firstly, it outperforms the previous best-performing distantly supervised method by a notable margin of 10.59 points in SARI Final (19.3% improvement). Secondly, even compared to ChatGPT prompted with in-context examples, LIFE achieves a superiority of 7.16 points in SARI Final. Copyright © 2023, The Authors. All rights reserved.

关键词： Error correction

来源：评论

学校读者我要写书评

暂无评论

Testing Non-Commutativity of Reduce Functions with Multi-Column Inputs

SSRN

引用

SSRN 2024年

作者： Zhang, Xuan Zhu, Chenlu Li, Ning Zhang, Peng Liu, Lei College of Software Jilin University Jilin China College of Computer Science and Technology Jilin University Jilin China Key Laboratory of Symbolic Computation and Knowledge Engineering Jilin University Jilin China

With the continuous development of the MapReduce programming model, it is necessary to ensure the reliability of MapReduce programs. In practice, the non-commutativity of Reduce functions seriously affects the reliability of the MapReduce program, which is difficult to debug and even causes errors. Current researches on the non-commutability detection of Reduce function consider the case that the input value is a single attribute. However, such researches ignore the situation where inputs to most reduce functions in practical applications consist of multiple columns (such as a table). To test the commutativity of reduce functions where each input record may contain several input attributes, a new testing method is proposed. This approach uses symbolic execution tools to help generate a few input records, and breaks their data dependencies to generate an original test case t0, with a dynamic program slicing technique to lessen the scale of t0. And the ultimate test suite is consisted of different permutations of records in t0. In the end, experiments demonstrate the effectiveness of our testing method and that the permutation method Gm is helpful to reduce its complexity. © 2024, The Authors. All rights reserved.

关键词： MapReduce

来源：评论

学校读者我要写书评

暂无评论

Learning a Typhoon Bayesian Network Structure from Natural Language Reports 5th

Learning a Typhoon Bayesian Network Structure from Natural...

引用

5th IFIP TC 12 International Conference on Intelligence Science, ICIS 2022

作者： Yao, Zhangrui Chen, Junhan Pan, Yinghui Zeng, Yifeng Ma, Biyang Ming, Zhong Department of Automation and Xiamen Key Laboratory of Big Data Intelligent Analysis and Decision-Making Xiamen University Xiamen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Department of Computer and Information Sciences Northumbria University Newcastle upon Tyne United Kingdom School of Computer Science and Engineering Minnan Normal University Zhangzhou China

ISBN: (纸本)9783031149023

Given the huge toll caused by natural disasters, it is critically important to develop an effective disaster management and emergency response technique. In this article, we investigate relationships between typhoon-related variables and emergency response from natural language (NL) reports. A major challenge is to exploit typhoon state information for typhoon contingency plan generation, especially from unstructured text data based on NL input. To tackle this issue, we propose a novel framework for learning typhoon Bayesian network structures (FLTB), which can extract typhoon state information from unstructured NL, mine inter-information causal relationships and then generate Bayesian networks. We first extract information about typhoon states through NL processing (NLP) techniques, and then analyze typhoon reports by designing heuristic rules to identify causal relationships between states. We leverage these features to improve the learned structures and provide user-interaction mechanisms to finalize Bayesian networks. We evaluate the performance of our framework on real-world typhoon datasets and develop the Bayesian networks based typhoon emergency response systems. © 2022, IFIP International Federation for Information Processing.

关键词： Bayesian networks

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Graph Convolutional Network for Visual Reinforcement Learning of Action Detection 8

Heterogeneous Graph Convolutional Network for Visual Reinfor...

引用

8th International Conference on Automation, Control and Robotics engineering, CACRE 2023

作者： Wang, Liangliang Huang, Chengxi Chen, Xinwei Wuhan University of Technology School of Computer Science and Artificial Intelligence Luoshi Road 122 Wuhan430070 China -HIT Research Institute of Engineering and Technology Building 9 Software Park Quanzhou362008 China Minjiang University Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Fuzhou350121 China

ISBN: (纸本)9798350302776

Existing action detection approaches do not take spatio-temporal structural relationships of action clips into account, which leads to a low applicability in real-world scenarios and can benefit detecting if exploited. To this end, this paper proposes to formulate the action detection problem as a reinforcement learning process which is rewarded by observing both the clip sampling and classification results via adjusting the detection schemes. In particular, our framework consists of a heterogeneous graph convolutional network to represent the spatio-temporal features capturing the inherent relation, a policy network which determines the probabilities of a predefined action sampling spaces, and a classification network for action clip recognition. We accomplish the network joint learning by considering the temporal intersection over union and Euclidean distance between detected clips and ground-truth. Experiments on ActivityNet v1.3 and THUMOS14 demonstrate our method. © 2023 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

ICTDroid: Parameter-Aware Combinatorial Testing for Components of Android Apps

ICTDroid: Parameter-Aware Combinatorial Testing for Componen...

引用

IEEE International Conference on Automated software engineering (ASE)

作者： Shixin Zhang Shanna Li Xi Deng Jiwei Yan Jun Yan Technology Center of Software Engineering Institute of Software Chinese Academy of Sciences Information Technology Center Beijing Jiaotong University State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences University of Chinese Academy of Sciences

Components are the fundamental building blocks of Android applications. Different functional modules represented by components often rely on inter-component communication mechanisms to achieve cross-module data transfer and method invocation. It is necessary to conduct robustness testing on components to prevent component launching crashes and privacy leaks caused by unexpected input parameters. However, as the complexity of the input parameter structure and the diversity of possible inputs, developers may overlook specific inputs that result in exceptions. At the same time, the vast input space also brings challenges to efficient component testing. In this paper, we designed an automated test generation and execution tool for Android application components named ICTDroid, which combines static parameter extraction and adaptive-strength combinatorial testing generation to detect bugs with a compact test suite. Experiments have shown that the tool triggers 205 unique exceptions in 30 open-source applications with 1,919 test cases in 83 minutes, where the developers have confirmed six defects in three issues we reported.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Detection of Java Basic Thread Misuses Based on Static Event Analysis

Detection of Java Basic Thread Misuses Based on Static Event...

引用

IEEE International Conference on Automated software engineering (ASE)

作者： Baoquan Cui Miaomiao Wang Chi Zhang Jiwei Yan Jun Yan Jian Zhang State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Technology Center of Software Engineering Institute of Software Chinese Academy of Sciences Beijing China

The fundamental asynchronous thread (***. Thread) in Java can be easily misused, due to the lack of deep understanding for garbage collection and thread interruption mechanism. For example, a careless implementation of asynchronous thread may cause no response to the interrupt mechanism in time, resulting in unexpected thread-related behaviors, especially resource leak/waste. Currently, few works aim at these misuses and related works adopt either the dynamic approach which lacks effective inputs or the static path-sensitive approach with high time consumption due to the path explosion, causing false negatives. We have found that the behavior of threads and the interaction between threads and its referencing objects can be abstracted. In this paper, we propose an event analysis approach to detect the defects in Java programs and Android apps, which focuses on the existence or the order of the events to reduce the false negatives. We extract the misuse-related events, containing the thread events and the destroy events of the object referenced by the thread. Then we analyze the events with loop identification, happens-before relationship construction and alias determination. Finally, we implement an automatic tool named Leopard and evaluate it on real world Java programs and Android apps. Experiments show that it is efficient when comparing with the existing approach (misuse: 723 vs 47, time: 60s vs 30min), which also outperforms the existing work in precision. The manual check indicates that Leopard is more efficient and effective than existing work. Besides, 66 issues reported by us have been confirmed and 21 of them have been fixed by developers.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Ant: An Efficient Lossless Compression Algorithm for IoT Time Series Data

Ant: An Efficient Lossless Compression Algorithm for IoT Tim...

引用

IEEE International Conference on Big Data and Cloud Computing (BdCloud)

作者： Junhui Li Guangping Xu Hongzhang Yang Yulei Wu School of Computer Science and Engineering Tianjin University of Technology Tianjin China Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology Tianjin China Key Laboratory of Computer Vision and System Ministry of Education China Faculty of Science and Engineering University of Bristol Bristol U.K

In various applications in Internet of Things like industrial monitoring, large amounts of floating-point time series data are generated at an unprecedented rate. Efficient compression algorithms can effectively reduce the size of data, enhance transmission performance and storage efficiency, and simultaneously lower storage costs. Therefore, there is a need for lightweight and efficient stream compression algorithms. In this paper, we propose a novel lossless floating-point data compression algorithm called Ant. The main idea is to encode double-precision floating-point numbers into integer form, calculate the delta between adjacent values, and then convert the delta into unsigned integers. This encoding method effectively reduces storage costs and improves data compression efficiency. Extensive experiments on real-world datasets demonstrate that our algorithm achieves compression speeds at least as fast as state-of-the-art streaming methods, and a 63% relative improvement in average compression rate.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Imperceptible Face Forgery Attack via Adversarial Semantic Mask

arXiv

引用

arXiv 2024年

作者： Liu, Decheng Su, Qixuan Peng, Chunlei Wang, Nannan Gao, Xinbo State Key Laboratory of Integrated Services Networks School of Cyber Engineering Xidian University Shaanxi Xi’an710071 China Shanghai Key Laboratory of Computer Software Evaluating and Testing Shanghai201112 China State Key Laboratory of Integrated Services Networks School of Telecommunications Engineering Xidian University Shaanxi Xi’an710071 China Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecommunications Chongqing400065 China

With the great development of generative model techniques, face forgery detection draws more and more attention in the related field. Researchers find that existing face forgery models are still vulnerable to adversarial examples with generated pixel perturbations in the global image. These generated adversarial samples still can't achieve satisfactory performance because of the high detectability. To address these problems, we propose an Adversarial Semantic Mask Attack framework (ASMA) which can generate adversarial examples with good transferability and invisibility. Specifically, we propose a novel adversarial semantic mask generative model, which can constrain generated perturbations in local semantic regions for good stealthiness. The designed adaptive semantic mask selection strategy can effectively leverage the class activation values of different semantic regions, and further ensure better attack transferability and stealthiness. Extensive experiments on the public face forgery dataset prove the proposed method achieves superior performance compared with several representative adversarial attack methods. The code is publicly available at https://***/clawerO-O/ASMA. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Recognizing Human-object Interactions for a Home Assistive Robot 8

Recognizing Human-object Interactions for a Home Assistive R...

引用

8th International Conference on Automation, Control and Robotics engineering, CACRE 2023

作者： Wang, Liangliang Huo, Guanglei Chen, Xinwei Huang, Chengxi Wuhan University of Technology School of Computer Science and Artificial Intelligence Luoshi Road 122 Wuhan430070 China -HIT Research Institute of Engineering and Technology Building 9 Software Park Quanzhou362008 China Minjiang University Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Fuzhou350121 China

ISBN: (纸本)9798350302776

To enable robots to understand a specific assistive task during human-robot interactions under complex home scenes, at the center is the problem of human-object interaction (HOI) recognition. In particular, aiming at identifying efficiently, this paper proposes a graph convolutional network (GCN) by considering the similarity and proximity of human and object nodes, based on which the geometric positions and morphological relationships are explored. Also, a supplementary convolutional neural network (CNN) is fused to GCN to extract the global semantic features, which also ensures the recognition applicability for the cases where detectors fall into failure. Experiments on V-COCO demonstrate our approach, and results in real-world scenes on our assistive robot - Xiaozhi II verify its robustness. © 2023 IEEE.

关键词： Human robot interaction

来源：评论

学校读者我要写书评

暂无评论

Deep Spatio-Temporal Decision Fusion Network for Facial Expression Recognition 1

引用

4th International Conference on Machine Learning for Cyber Security, ML4CS 2022

作者： Chen, Xuanchi Yang, Heng Zhang, Xia Zheng, Xiangwei Li, Wei School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan China Internet Diagnosis and Treatment Center Taian City Central Hospital Taian China Shandong Normal University Library Shandong Normal University Jinan China

ISBN: (数字)9783031201028

ISBN: (纸本)9783031201011

Facial expression recognition (FER) becomes research focus in affective computing, as it already plays an important role in public security application scenarios such as urban safety management and safety driving assistance systems. Modeling the spatiotemporal information of facial expression sequences in a targeted manner, integrating and utilizing them appropriately is challenging. In this paper, a facial expression recognition method based on spatial-temporal decision fusion network (STDFN) is proposed. Firstly, the facial expression sequences are divided into four sub-sequences according to face regions, and BiLSTM are used for each of sub-sequences to extract local temporal features. The local morphological features of facial expressions can be captured in more detail to maximize the utilization of the temporal features of dynamic facial expressions. Then, VGG19 is utilized to extract the shallow spatial features of peak expression frame, and the channel weights of spatial features is assigned by squeeze-and-excitation module to attain the weighted spatial features. This allows valid spatial features to be purposefully retained to avoid overfitting. Finally, temporal features and spatial features are used separately calculating expression classification results. And a decision-level fusion module is designed to fuse the two results to obtain the final FER result. Extensive experimental results demonstrate that on three FER datasets CK+, Oulu-CASIA and MMI, achieves 98.83%, 89.31% and 82.86% accuracy, which proved that STDFN effectively improved the recognition accuracy of FER. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Face recognition

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：