检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Xu, Fangzhi Lin, Qika Han, Jiawei Zhao, Tianzhe Liu, Jun Cambria, Erik The School of Computer Science and Technology Xi’an Jiao-tong University Shaanxi Xi’an710049 China The Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering National Engineering Lab for Big Data Analytics Shaanxi Xi’an710049 China The School of Computer Science and Engineering Nanyang Technological University Singapore

Logical reasoning consistently plays a fundamental and significant role in the domains of knowledge engineering and artificial intelligence. Recently, Large Language Models (LLMs) have emerged as a noteworthy innovation in natural language processing (NLP). However, the question of whether LLMs can effectively address the task of logical reasoning, which requires gradual cognitive inference similar to human intelligence, remains unanswered. To this end, we aim to bridge this gap and provide comprehensive evaluations in this paper. Firstly, to offer systematic evaluations, we select fifteen typical logical reasoning datasets and organize them into deductive, inductive, abductive and mixed-form reasoning settings. Considering the comprehensiveness of evaluations, we include 3 early-era representative LLMs and 4 trending LLMs. Secondly, different from previous evaluations relying only on simple metrics (e.g., accuracy), we propose fine-level evaluations in objective and subjective manners, covering both answers and explanations, including answer correctness, explain correctness, explain completeness and explain redundancy. Additionally, to uncover the logical flaws of LLMs, problematic cases will be attributed to five error types from two dimensions, i.e., evidence selection process and reasoning process. Thirdly, to avoid the influences of knowledge bias and concentrate purely on benchmarking the logical reasoning capability of LLMs, we propose a new dataset with neutral content. Based on the in-depth evaluations, this paper finally forms a general evaluation scheme of logical reasoning capability from six dimensions (i.e., Correct, Rigorous, Self-aware, Active, Oriented and No hallucination). It reflects the pros and cons of LLMs and gives guiding directions for future works. © 2023, CC BY.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Potential friendship discovery in social networks based on hybrid ensemble multiple collaborative filtering models in a 5G network environment

引用

Digital Communications and Networks 2022年第6期8卷 868-876页

作者： Hexuan Hu Zhenzhou Lin Qiang Hu Ye Zhang Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai UniversityNanjingChina College of Computer and Information Hohai UniversityNanjingChina Office of Teaching Affairs Nanjing University of Finance and EconomicsNanjingChina Electric Engineering College Tibet Agriculture&Animal Husbandry UniversityTibetChina

At present, 5G network technology is being applied to various social network modes, and it can provide technical and traffic support for social networks. Potential friendship discovery technology in 5G-enabled social networks is beneficial for users to make potential friends and expand their range of activities and social hierarchy, which is highly sought after in today's social networks and has great economic and application value. However, the sparsity of the dominant user association dataset in 5G-enabled social networks and the limitations of traditional collaborative filtering algorithms are two major challenges for the friend recommendation problem. Therefore, in order to overcome these problems regarding previous models, we propose a Hybrid Ensemble Multiple Collaborative Filtering Model (HEMCF) for discovering potential buddy relationships. The HEMCF model draws on a special autoencoder method that can effectively exploit the association matrix between friends and additional information to extract a hidden representation of users containing global structural information. Then, it uses the random walk-based graph embedding algorithm DeepWalk to extract another hidden representation of users in the buddy network containing local structural information. Finally, in the output module, the HEMCF model stacks and multiplies the two types of hidden representations of users to ensure that the information mentioned above is concentrated in the final output to generate the final prediction value. The magnitude of the prediction value represents the probability of the users being friends, with larger values representing a high probability of the two users being friends, and vice versa. Experimental results show that the proposed method boosts the accuracy of the relationship prediction over baselines on 3 real-world public datasets dramatically.

关键词： 5G network Social network Collaborative filtering Recommendation system Friendship discovering

来源：评论

学校读者我要写书评

暂无评论

MalPacDetector: An LLM-Based Malicious NPM Package Detector

引用

IEEE Transactions on Information Forensics and Security 2025年 20卷 6279-6291页

作者： Jian Wang Zhen Li Jixiang Qu Deqing Zou Shouhuai Xu Ziteng Xu Zhenwei Wang Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Laboratory Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Jinyinhu Laboratory Huazhong University of Science and Technology Wuhan China Jinyinhu Laboratory Wuhan China Department of Computer Science Laboratory for Cybersecurity Dynamics University of Colorado at Colorado Springs Colorado Springs CO USA Ant Technology Group Company Ltd. Hangzhou China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Laboratory Cluster and Grid Computing Laboratory School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

The Node Package Manager ( npm) registry contains millions of JavaScript packages widely shared between worldwide developers. However, npm has also been abused by attackers to spread malicious packages, highlighting the importance of detecting malicious npm packages. Existing malicious npm package detectors suffer from, among other things, high false positives and/or high false negatives. In this paper, we propose a novel Malicious npm Package Detector (MalPacDetector), which leverages Large Language Model (LLM) to automatically and dynamically generate features (rather than asking experts to manually define them). To evaluate the effectiveness of MalPacDetector and existing detectors, we construct a new npm package dataset, which overcomes the weaknesses of existing datasets (e.g., a small number of examples and a high repetition rate of malicious fragments). The experimental results show that MalPacDetector outperforms existing detectors by achieving a false positive rate of 1.3% and a false negative rate of 7.5%. In particular, MalPacDetector detects 39 previously unknown malicious packages, which are confirmed by the npm security team.

关键词： Feature extraction Detectors Malware Codes Training Security data mining Large language models big data Syntactics

来源：评论

学校读者我要写书评

暂无评论

A Weighted Flat Lattice Transformer-based knowledge Extraction Architecture for Chinese Named Entity Recognition

A Weighted Flat Lattice Transformer-based Knowledge Extracti...

引用

International Conference on Computer Supported Cooperative Work in Design

作者： Hengwei Zhang Yuejia Wu Jian-Tao Zhou Inner Mongolia University College of Computer Science Hohhot China Ministry of Education National & Local Joint Engineering Research Center of Intelligent Information Processing Technology for Mongolian Engineering Research Center of Ecological Big Data Hohhot China Inner Mongolia Engineering Laboratory for Cloud Computing and Service Software Hohhot China Inner Mongolia Key Laboratory of Social Computing and Data Processing Hohhot China Inner Mongolia Key Laboratory of Discipline Inspection and Supervision Big Data Hohhot China Inner Mongolia Engineering Laboratory for Big Data Analysis Technology Hohhot China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

Named Entity Recognition (NER) is one of the contents of knowledge Extraction (KE) that transforms data into knowledge representation. However, Chinese NER faces the problem of lacking clear word boundaries that limit the effectiveness of the KE. Although the flat lattice Transformer (FLAT) framework, which converts lattice structure into a flat structure including a set of spans, can effectively improve this problem and obtain advanced results, there still exist the problems of insensitivity to entity importance weights and insufficient feature learning. This paper proposes a weighted flat lattice Transformer architecture for Chinese NER, namely WFLAT. The WFLAT first adds a weight matrix into self-attention calculation, which can obtain finer-grained partitioning of entities to improve experimental performance, and then adopts a multi-layer Transformer encoder with each layer using a multi-head self-attention mechanism. Extensive experimental results on benchmarks demonstrate that our proposed KE model can obtain state-of-the-art performance for the Chinese NER task.

关键词： Representation learning Computational modeling Memory management Lattices Named entity recognition Transforms knowledge representation

来源：评论

学校读者我要写书评

暂无评论

Experimentally Study on the Effect of RIE Etching Power on Etching Rate of β-Ga2O3Thin Film 29

Experimentally Study on the Effect of RIE Etching Power on E...

引用

29th International Symposium on Semiconductor Manufacturing, ISSM 2022

作者： Xu, Wang Yang, Ran Jing Lai, Yang Shun, Yang Fa Kui, Ma College of Big Data and Information Engineering Guizhou University China Guizhou Provincial Key Laboratory of Micro and Nano Electronics and Software Technology Guizhou Province Guiyang City China

ISBN: (纸本)9781665471336

As a member of ultra-wide band gap semiconductor materials, β-Ga2O3 materials have attracted wide attention from researchers in the semiconductor field in recent years. Etching process is crucial to realize semiconductor devices and integrated circuits based on β-Ga2O3 materials. Based on the reaction ion etching process commonly used in silicon-based semiconductor technology, the etching experiment research of β-Ga2O3 thin film is carried out. The β-Ga2O3film is etched with SF6, based on the induction coupled reaction ion etching. The effect of RIE etching power, excitation power and bias power, on etching rate of β-Ga2O3 thin film has been studied. SEM characterization results show that the etching rate is the highest at 600W excitation power. The etching rate increases with the increase of bias power. The etching rate at 200W bias power is slightly higher than that at 150W bias power. However, the photoresist used as the etch mask will be damaged at 200W bias power. © 2022 IEEE.

关键词： Etching

来源：评论

学校读者我要写书评

暂无评论

Signage Detection Based on Adaptive SIFT 1

引用

11th International Conference on Frontiers of Intelligent Computing: Theory and Applications, FICTA 2023

作者： Wang, Jiaji Chen, Shuwen Shao, Jiaqi Qu, Hui Wang, Ziyi School of Physics and Information Engineering Jiangsu Second Normal University Nanjing211200 China Jiangsu Province Engineering Research Center of Basic Education Big Data Application Nanjing211200 China State Key Laboratory of Millimeter Waves Southeast University Nanjing210096 China

ISBN: (数字)9789819967063

ISBN: (纸本)9789819967056

In navigation and wayfinding applications, signage is crucial for finding destinations. This paper proposes a new method for detecting signage, with the aim of helping blind individuals navigate unfamiliar indoor environments. To enhance the accuracy of signage detection and eliminate interference, the proposed method initially obtains the attended areas through a saliency map. The signage is then identified in these attended areas using SIFT algorithm-based adaptive threshold. This innovative approach is capable of detecting and recognizing multiple signs. Finally, audio facilitates the detection and recognition of results transferred to the users. Our experimental results, based on indoor signage datasets we collected, demonstrate the efficiency of our proposed method. On average, it takes 0.104 s to process the data. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Computer programming

来源：评论

学校读者我要写书评

暂无评论

Convolution Neural Network for Renal Function Assessment Based on Glomerular Filtration Rate

Convolution Neural Network for Renal Function Assessment Bas...

引用

2021 International Conference on Advanced Technologies and Applications of Modern Industry, ATAMI 2021

作者： Wu, Congzhong Wang, Lulu Yang, Zhen Wang, Jing Han, Li Li, Fei Key Laboratory of Knowledge Engineering with Big Data Ministry of Education Hefei University of Technology Hefei Anhui230601 China School of Computer and Information Hefei University of Technology Hefei Anhui230601 China The Second Hospital of Anhui Medical University Hefei Anhui230601 China

Objective: To develop a Convolution Neural Network (CNN) and discuss its performance in the estimation of Glomerular Fifiltration Rate (GFR) for patients with chronic kidney disease (CKD). Methods: A total of 112 patients with chronic kidney disease were enrolled in this study. The GFR is measured by 99mTc-DTPA renal dynamic and used as standard GFR after normalization by body surface area imaging. We established a CNN model and verified the performance of the model by comparing the GFR predicted by the model with the standard GFR. It turned out that the CNN could better evaluate the GFR of patients which is superior to the CG formula, MDRD formula, CKD-EPI formula and GRNN model. Conclusions: The CNN significantly evaluated GFR for patients with CKD, and it showed better performance than traditional methods and GRNN model and closer to the results of 99mTc-DTPA. The experimental results demonstrated that CNN could be used to estimate GFR. © 2022 Institute of Physics Publishing. All rights reserved.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Research on Rice Disease Identification Based on Improved ShuffleNet V2 Network 4

Research on Rice Disease Identification Based on Improved Sh...

引用

4th International Symposium on Computer Technology and Information Science, ISCTIS 2024

作者： Chen, Zhihuan Gu, Shenming Yu, Minghui Huang, Zhaoquan Zhejiang Ocean University Key Laboratory of Oceanographic Big Data Mining & Application of Zhejiang Province Zhoushan310622 China Ningbo University School of Faculty of Electrical Engineering and Computer Science Ningbo315000 China Zhejiang Normal University College of Mathematics and Computer Science Jinhua321000 China

ISBN: (纸本)9798350354560

How to efficiently and accurately identify rice diseases plays a crucial role in rice planting and production. Deploying large convolutional neural networks in practical applications is challenging due to their complex model parameters, while lightweight networks often compromise accuracy to reduce model size. In this study, we introduce a lightweight and precise network called GM-ShuffleNet for rice disease recognition. Building upon the ShuffleNet V2 network, we integrated the Ghost module and MCA attention mechanism, adjusted the network structure and number of channels, and replaced the ReLU activation function with Mish. Through comparative experiments, the accuracy of GM-ShuffleNet algorithm proposed in this paper reaches 96.20%, which is superior to ShuffleNet V2 and other excellent lightweight networks. In addition, in terms of model complexity comparison, the number of model parameters and FLOPs decreased by 82.54% and 77.44% respectively compared with the original network. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Intelligent Aquaculture Water Quality Detection Scheme Based on TurMass Technology 8

Intelligent Aquaculture Water Quality Detection Scheme Based...

引用

8th International Conference on Intelligent Computing and Signal Processing, ICSP 2023

作者： Zhang, Yafang Huang, Yongxin Chen, Hongming Zhejiang Ocean University School of Information Engineering Zhoushan316022 China Zhejiang Ocean University Key Laboratory of Oceanographic Big Data Mining & Application of Zhejiang Province Zhoushan316022 China Nankai University College of Electronics Information and Optical Engineering Tianjin300350 China

ISBN: (纸本)9798350302455

To address the shortcomings of wireless data transmission in aquaculture water quality monitoring, we have developed an intelligent water quality detection device based on the STM32 microcontroller system as the processing unit and TurMass transmission technology. This device can detect water quality in real-time and transmit the data to the user end for real-time monitoring so that users can monitor water quality in real-time. Through rigorous and systematic experimental measurements, the accuracy of the detected data by the device was found to be extremely high, indicating that it can be used in practical applications. © 2023 IEEE.

关键词： Water quality

来源：评论

学校读者我要写书评

暂无评论

Contamination-Resilient Anomaly Detection via Adversarial Learning on Partially-Observed Normal and Anomalous data 41

Contamination-Resilient Anomaly Detection via Adversarial Le...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Lv, Wenxi Su, Qinliang Wan, Hai Xu, Hongteng Xu, Wenchao School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China Gaoling School of Artifical Intelligence Renmin University of China Beijing China Department of Computing The Hong Kong Polytechnic University Hong Kong

Many existing anomaly detection methods assume the availability of a large-scale normal dataset. But for many applications, limited by resources, removing all anomalous samples from a large unlabeled dataset is unrealistic, resulting in contaminated datasets. To detect anomalies accurately under such scenarios, from the probabilistic perspective, the key question becomes how to learn the normal-data distribution from a contaminated dataset. To this end, we propose to collect two additional small datasets that are comprised of partially-observed normal and anomaly samples, and then use them to help learn the distribution under an adversarial learning scheme. We prove that under some mild conditions, the proposed method is able to learn the correct normal-data distribution. Then, we consider the overfitting issue caused by the small size of the two additional datasets, and a correctness-guaranteed flipping mechanism is further developed to alleviate it. Theoretical results under incomplete observed anomaly types are also presented. Extensive experimental results demonstrate that our method outperforms representative baselines when detecting anomalies under contaminated datasets. Copyright 2024 by the author(s)

关键词： Adversarial machine learning

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：