检索结果-内蒙古大学图书馆

4th International Signal Processing, Communications and Engineering Management Conference, ISPCEM 2024

作者： Chunyan, Wang Kaixin, Fu Xiang, Wang School of Software Liaoning Technical University Huludao China Department of Computer Science and Technology Dalian University of Technology Dalian China

ISBN: (纸本)9798331528676

To address the increased complexity of surface cover, increased heterogeneity within homogeneous regions, and increased similarity between different regions in high-resolution remote sensing images, which lead to increased classification difficulty, a supervised learning method based on a double neighborhood relationship Gaussian Regression Mixture Model (GRMM) is proposed. Firstly, supervised sampling of image regions is conducted, fitting histograms using the least squares method to establish Gaussian Mixture Models (GMM) for each land cover to represent the complex gray-scale features of land cover. Secondly, local spatial information of adjacent pixels is introduced into the image gray-scale space to construct a Gaussian Regression Model (GRM). Finally, in the membership degree space, the neighborhood relationship is processed again to achieve classification decision. The kappa coefficients of GRMM on synthetic images and real high-resolution remote sensing images respectively reaches 97.2% and 98.5%. Compared to existing mainstream models, GRMM demonstrates strong classification efficiency, noise reduction capability, and generalization ability, with clear classification boundaries, effectively improved the land cover classification capability of remote sensing images. ©2024 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Scale adaptive fitness evaluation‐based particle swarm optimisation for hyperparameter and architecture optimisation in neural networks and deep learning

引用

CAAI Transactions on Intelligence technology 2023年第3期8卷 849-862页

作者： Ye‐Qun Wang Jian‐Yu Li Chun‐Hua Chen Jun Zhang Zhi‐Hui Zhan School of Computer Science and Engineering South China University of TechnologyGuangzhouChina School of Software Engineering South China University of TechnologyGuangzhouChina Hanyang University AnsanSouth Korea

Research into automatically searching for an optimal neural network(NN)by optimi-sation algorithms is a significant research topic in deep learning and artificial ***,this is still challenging due to two issues:Both the hyperparameter and ar-chitecture should be optimised and the optimisation process is computationally *** tackle these two issues,this paper focusses on solving the hyperparameter and architecture optimization problem for the NN and proposes a novel light‐weight scale‐adaptive fitness evaluation‐based particle swarm optimisation(SAFE‐PSO)***,the SAFE‐PSO algorithm considers the hyperparameters and architectures together in the optimisation problem and therefore can find their optimal combination for the globally best ***,the computational cost can be reduced by using multi‐scale accuracy evaluation methods to evaluate ***,a stagnation‐based switch strategy is proposed to adaptively switch different evaluation methods to better balance the search performance and computational *** SAFE‐PSO algorithm is tested on two widely used datasets:The 10‐category(i.e.,CIFAR10)and the 100−cate-gory(i.e.,CIFAR100).The experimental results show that SAFE‐PSO is very effective and efficient,which can not only find a promising NN automatically but also find a better NN than compared algorithms at the same computational cost.

关键词： deep learning evolutionary computation hyperparameter and architecture optimisation neural networks particle swarm optimisation scale‐adaptive fitness evaluation

来源：评论

学校读者我要写书评

暂无评论

CATC: A task scheduling based on computing-aware and task classification in Pervasive Edge Computing 22

CATC: A task scheduling based on computing-aware and task cl...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Yang, JiaHui Zhang, Yang Zhang, ShuKui Zhu, Mingyu Wang, YingYing Soochow University School of Computer Science and Technology Suzhou215006 China Global Institute of Software Technology Suzhou215163 China

ISBN: (纸本)9798331509712

With no infrastructure, the battlefield environment has characteristics such as device heterogeneity, uneven resource distribution, and dynamic node movement. Additionally, the sensing tasks produced by devices are complex and delay-sensitive, making it challenging to complete the tasks within the time limit if a centralized unified scheduling strategy is employed. In order to fully utilize the resources in the battlefield environment, optimize the load situation in the battlefield scenario, and guarantee the stability of task offloading, this paper proposes a scheduling strategy based on computing-aware and task classification (CATC). We first design a distributed computing-aware mechanism to ensure the detection of surrounding resources and load conditions by the generating task devices. Based on this, we classify the battlefield tasks into multi-attribute tasks according to the actual situation, and develop an offloading strategy for different task types, taking into account computation, storage, transmission, and load situation. The simulation results demonstrate that the proposed CATC strategy has a high task completion rate and can satisfy the definite delay requirements of delay-sensitive tasks. © 2024 IEEE.

关键词： Computation offloading

来源：评论

学校读者我要写书评

暂无评论

Fix the Tests: Augmenting LLMs to Repair Test Cases with Static Collector and Neural Reranker 35

Fix the Tests: Augmenting LLMs to Repair Test Cases with Sta...

引用

35th IEEE International Symposium on software Reliability Engineering, ISSRE 2024

作者： Liu, Jun Yan, Jiwei Xie, Yuanyuan Yan, Jun Zhang, Jian State Key Laboratory of Computer Science Institute of Software China Chinese Academy of Sciences Technology Center of Software Engineering Institute of Software China China UCAS School of Intelligent Science and Technology Hangzhou Institute for Advanced Study China

ISBN: (纸本)9798350353884

During software evolution, it is advocated that test code should co-evolve with production code. In real development scenarios, test updating may lag behind production code changing, which may cause compilation failure or bring other troubles. Existing techniques based on pre-trained language models can be directly adopted to repair obsolete tests caused by such unsynchronized code changes, especially syntactic-related ones. However, the lack of task-oriented contextual information affects the repair accuracy on large-scale projects. Starting from an obsolete test, the key challenging task is precisely identifying and constructing Test-Repair-Oriented Contexts (TROCtxs) from the whole repository within a limited token *** this paper, we propose Synter (SYNtactic-breaking- changes-induced TEst Repair), a novel approach based on LLMs to automatically repair obsolete test cases via precise and concise TROCtxs construction. Inspired by developers' programming practices, we design three types of TROCtx: class context, usage context, and environment context. Given an obsolete test case to repair, Synter firstly collects the related code information for each type of TROCtx through static analysis techniques automatically. Then, it generates reranking queries to identify the most relevant TROCtxs, which will be taken as the repair-required key contexts and be input to the large language model for the final test *** evaluate the effectiveness of Synter, we construct a benchmark dataset that contains a set of obsolete tests caused by syntactic breaking changes. The experimental results show that Synter outperforms baseline approaches both on textual- and intent-matching metrics. With the augmentation of constructed TROCtxs, hallucinations are reduced by 57.1%. © 2024 IEEE.

关键词： Syntactics

来源：评论

学校读者我要写书评

暂无评论

A Steel Surface Defect Detection Model Based on YOLOv7-tiny

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer Science 2024年第12期51卷 2074-2082页

作者： Wang, Yayun Tao, Ye Cui, Wenhua Shen, Lijia Anshan Normal University Anshan China School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan China School of Electronic and Information Engineering University of Science and Technology Liaoning Anshan China

This article designs the PELAN structure based on the lightweight YOLOv7-tiny model for surface defect detection of hot-rolled steel strips. At the same time, the CA (Channel Attention) is embedded in the feature pyramid structure and the Slim-Neck structure is introduced at the neck. Through experimental result analysis, it is proved that the lightweight algorithm not only reduces model complexity, but also has better detection accuracy and processing speed. This article focuses on exploring the potential of lightweight YOLOv7-tiny models in steel surface defect detection. With the advancement of Industry 4.0, quickly and accurately detecting surface defects on steel strips has become crucial. However, traditional defect detection methods often face issues such as large computational requirements and poor real-time performance. Therefore, we have designed a novel PELAN structure that combines CA(Channel Attention)and Slim-Neck structure. The aim is to reduce model complexity while maintaining or even improving detection accuracy and processing speed. © (2024), (International Association of Engineers). All rights reserved.

关键词： Hot rolling

来源：评论

学校读者我要写书评

暂无评论

Robust Stability Region of PI Controllers for an Interval Time Delayed Load Frequency Control

Robust Stability Region of PI Controllers for an Interval Ti...

引用

2024 IEEE PES Innovative Smart Grid Technologies Europe Conference, ISGT EUROPE 2024

作者： Ghorbani, Majid Nosrati, Komeil Tepljakov, Aleksei Belikov, Juri Petlenkov, Eduard Tallinn University of Technology The Department of Computer Systems Tallinn Estonia Tallinn University of Technology Department of Software Science Tallinn Estonia

ISBN: (纸本)9789531842976

Transmission delays, load fluctuations, and intermittent output power of renewable energy sources will all have a substantial impact on frequency stability in microgrids (MGs). In response, various types of advanced control techniques were employed to improve the performance of the load frequency control (LFC) systems. Notwithstanding this, the fragility of the current techniques makes LFC systems vulnerable to parametric (interval) uncertainties associated with communication time delay. Inspired by recent breakthroughs in robust control, in this study, we device a novel graphical tuning method of the PI controller for tackling this issue. First, we determine the shape of the value set for the characteristic function of uncertain LFC system. Then, the exposed edges are constructed to the boundary of this set under interval uncertainties. To accomplish this step and thanks to the D-decomposition method, the robust stability region in the parameter space of the controllers are determined. Our devised methodology contributes a family of robust stabilizing controllers, instead of just one, where it enables a more flexible choice for the controller parameters. The efficacy of the proposed methodology is indicated on a given single-area MG. © 2024 IEEE.

关键词： Proportional control systems

来源：评论

学校读者我要写书评

暂无评论

Unleashing the Power of LLMs in Court View Generation by Stimulating Internal Knowledge and Incorporating External Knowledge

Unleashing the Power of LLMs in Court View Generation by Sti...

引用

2024 Findings of the Association for Computational Linguistics: NAACL 2024

作者： Liu, Yifei Wu, Yiquan Li, Ang Zhang, Yating Sun, Changlong Lu, Weiming Wu, Fei Kuang, Kun School of Software Technology Zhejiang University China College of Computer Science and Technology Zhejiang University China Alibaba Group China

ISBN: (纸本)9798891761193

Court View Generation (CVG) plays a vital role in the realm of legal artificial intelligence, which aims to support judges in crafting legal judgment documents. The court view consists of three essential judgment parts: the charge-related, law article-related, and prison term-related parts, each requiring specialized legal knowledge, rendering CVG a challenging task. Although Large Language Models (LLMs) have made remarkable strides in language generation, they encounter difficulties in the knowledge-intensive legal domain. Actually, there can be two types of knowledge: internal knowledge stored within LLMs' parameters and external knowledge sourced from legal documents outside the models. In this paper, we decompose court views into different parts, stimulate internal knowledge, and incorporate external information to unleash the power of LLMs in the CVG task. To validate our method, we conduct a series of experiment results on two real-world datasets LAIC2021 and CJO2022. The experiments demonstrate that our method is capable of generating more accurate and reliable court views. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

COLARE: Commit Classification via Fine-grained Context-aware Representation of Code Changes 31

COLARE: Commit Classification via Fine-grained Context-aware...

引用

31st IEEE International Conference on software Analysis, Evolution and Reengineering, SANER 2024

作者： Zeng, Qunhong Zhang, Yuxia Sun, Zeyu Guo, Yujie Liu, Hui School of Computer Science & Technology Beijing Institute of Technology Beijing China Institute of Software Chinese Academy of Sciences Beijing China

ISBN: (纸本)9798350330663

Commit classification for maintenance activities is of critical importance for both industry and academia. State-of-the-art approaches either treat code changes as plain text or rely on manually identified features. Directly applying the most advanced model of code change representation into commit classification faces two limitations: (1) coarse-grained diff comparison neglects the distance of modified code lines;(2) missing key context information of hunk modification and file categories. This study proposes a novel classification model, COLARE, which compares code changes at the hunk level, takes fine-grained features based on categories of changed files, and aggregates with the representation of commit messages. The evaluation results show that our model outperforms state-of-the-art techniques by 7.24% and 7.35% in accuracy and macro F1 score, respectively. We also manually labeled a multi-language dataset and evaluated our approach, The results further confirm that our approach achieves the best performance over three baselines, including ChatGPT (3.5). The evaluation of the ablation study demonstrates the effectiveness of the major components in our technique. © 2024 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

UNiFY-ESSL: UNIfied Framework for Yielding EEG-Based Emotion Recognition Model with Self-Supervised Learning

UNiFY-ESSL: UNIfied Framework for Yielding EEG-Based Emotion...

引用

2024 International Conference on Brain computer Interface and Healthcare Technologies, iCon-BCIHT 2024

作者： Ahuja, Chirag Sethia, Divyashikha Delhi Technology University Department of Computer Engineering Delhi India Delhi Technology University Department of Software Engineering Delhi India

ISBN: (纸本)9798331540067

Electroencephalography (EEG) based emotion recognition shows promise in human-computer interaction and mental health monitoring, but faces challenges in cross-dataset generalization. This study introduces the Unified Framework for EEG-based Emotion Recognition with Self-Supervised Learning (UNiFY-ESSL), combining self-supervised learning (SSL) and advanced sampling techniques to integrate multiple EEG datasets. UNiFY-ESSL utilizes a unified 14-channel setup across all datasets (SEED, DEAP, and DREAMER), allowing consistent preprocessing and feature extraction. This approach adapts datasets with varying initial channel counts (62 for SEED, 32 for DEAP, and 14 for DREAMER) to a common configuration based on DREAMER's channels. The framework explores contrastive learning methods, specifically Simple Contrastive Learning (Sim-CLR) and Contrastive Predictive Coding (CPC), enhanced by a novel stratified sampling scheme. Experiments yield impressive results: SimCLR achieves F1-scores of 82.62%, 87.83%, and 89.05% for SEED, DEAP, and DREAMER respectively, while CPC attains 81.35%, 82.27%, and 91.23%. It demonstrates improved cross-dataset generalization, with a 1-2% performance gain on DREAMER and maintained performance on DEAP despite channel reduction. However, SEED experiences a 3% F1-score drop due to significant channel reduction. These findings highlight the potential of our sampling method in advancing EEG-based emotion recognition while underscoring the need for future work on efficient channel combination strategies, particularly for datasets with higher initial channel counts. © 2024 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

DROWN: Towards Tighter LiRPA-based Robustness Certification 31

DROWN: Towards Tighter LiRPA-based Robustness Certification

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Zhang, Yunruo Du, Tianyu Ji, Shouling Guo, Shanqing School of Cyber Science and Technology Shandong University China School of Software Technology Zhejiang University China College of Computer Science and Technology Zhejiang University China

ISBN: (纸本)9798891761964

The susceptibility of deep neural networks to adversarial attacks is a well-established concern. To address this problem, robustness certification is proposed, which, unfortunately, suffers from precision or scalability issues. In this paper, we present DROWN (Dual CROWN), a novel method for certifying DNNs' robustness. The advantage of DROWN is that it tightens classic LiRPA-based methods yet maintains similar scalability, which comes from refining pre-activation bounds of ReLU (and other activations) relaxations using two pairs of linear bounds derived from different relaxations of ReLU units in previous layers. The extensive evaluations show that DROWN achieves up to 83.39% higher certified robust accuracy than the baseline on CNNs and up to 4.68 times larger certified radii than the baseline on Transformers. Meanwhile, the running time of DROWN is about twice that of the baseline. © 2025 Association for Computational Linguistics.

关键词： Deep neural networks

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：