检索结果-内蒙古大学图书馆

Deep ensemble model for skin cancer classification with improved feature set

Multimedia Tools and Applications 2025年第10期84卷 7599-7626页

作者： Burada, Sreedhar Manjunathswamy, B.E. Kumar, M. Sunil Department of Computer Science and Engineering Don Bosco Institute of Technology Karnataka Bengaluru560074 India Department of Computer Science and Engineering School of Computing Mohan Babu University Sree Vidyanikethan Engineering College Andhra Pradesh Tirupati517102 India

Early identification of skin cancer is mandatory to minimize the worldwide death rate as this disease is covering more than 30% of mortality rates in young and adults. Researchers are in the move of proposing advanced models for the accurate classification of skin disease. This paper introduces a novel skin cancer detection model under the stages of preprocessing, feature extraction, classification as well as segmentation. The given input is first preprocessed via Gaussian filtering. Subsequently, the Deep learning model termed as ‘U-net’ is used for the segmentation process. Afterward, the extraction of the feature set including features like Improved Center symmetric Local Binary Pattern (ICS-LBP), Local Gradient Pattern (LGP), Median binary pattern (MBP) is performed. Classification takes place using Deep Ensemble Model based on the extracted features, that includes BiGRU, Deep Maxout, and Improved Deep Convolutional Neural Network (CNN), respectively. Also, as the hyperparameter tuning of the classification model enhances the performance, this work introduces a new Hybrid Atom and Arithmetic Operation Algorithm (HAAOA) algorithm for tuning the optimal weights of the Bi-GRU (Bi-directional Gated Recurrent Unit) and Deep Maxout model. At last, the validation of the proposed work is performed through different performance measures by comparing it with different algorithms. Accordingly, accuracy of the adopted method at 90th learning percentage is 93.56%, which is (~) 82.63%, 85.15%, 84.98%, 83.66% and 83.42% better than AOA, ASO, BWO, BMO and DO, respectively. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Improved Moth Flame Optimization Based Cluster Head Selection and Data Aggregation Using Machine Learning Approach

Improved Moth Flame Optimization Based Cluster Head Selectio...

引用

International Conference on 6G Communications Networking and Signal Processing, SGCNSP 2023

作者： Nadhan, Archana S Shreenath, K.N. Ramadan, Ghazi Mohamad Chanti, Yerrolla Bhukya, Shankar Nayak Department of Computer Science and Engineering GITAM School of Technology GITAM Deemed to Be University Bengaluru India Department of CSE Siddaganga Institute of Technology Tumakuru India College of MLT Ahl Al Bayt University Karbala Iraq School of Computer Science and Artificial Intelligence SR University Warangal India CMR Technical Campus Hyderabad Telangana Hyderabad India

ISBN: (纸本)9789819778751

Wireless Sensor Network is comprised with group of sensor nodes which are utilized in various field of applications. These sensor nodes are comprised with minimal processing ability due to diminished battery power. Since WSNs are vulnerable to failure because of power issues, the data aggregation plays a major role in WSN. The majority of the power is wasted due to redundant data from sensor nodes to base station. So, this research introduced an effective data aggregation scheme using Support Vector Machine (SVM) and the selection of CH takes place using the proposed Improved Moth Flame Optimization Algorithm (IMFO). The experimental results show that energy consumption of proposed IMFO for 200 nodes is 89.56 J whereas the existing Cluster based Reliable Data aggregation CRDA consumed 95.45 J. Similarly, the PDR of the proposed approach for 20 nodes is 0.83% whereas the existing Sail Fish Optimization with Support Vector Machine (SFO-SVM) achieved throughput of 0.71%. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Data aggregation

来源：评论

学校读者我要写书评

暂无评论

Toward actionable testing of deep learning models

引用

science China(Information sciences) 2023年第7期66卷 300-302页

作者： Yingfei XIONG Yongqiang TIAN Yepang LIU Shing-Chi CHEUNG Key Laboratory of High Confidence Software Technologies(Peking University) Ministry of Education School of Computer Science Peking University Department of Computer Science and Engineering The Hong Kong University of Science and Technology School of Computer Science University of Waterloo Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Department of Computer Science and Engineering Southern University of Science and Technology

Deep learning has become an important computational paradigm in our daily lives with a wide range of applications,from authentication using facial recognition to autonomous driving in smart vehicles. The quality of the deep learning models, i.e., neural architectures with parameters trained over a dataset, is crucial to our daily living and economy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Developing a Consumer Electronics Robotics With a Large Language Model Based on a Trustworthy AI Framework

引用

IEEE Transactions on Consumer Electronics 2025年第1期71卷 2027-2038页

作者： Wu, Hsin-Te Wei, Wei Li, Syuan-Hao Chen, Mu-Yen National Taitung University Computer Science and Information Engineering Taitung950309 Taiwan Xi'an University of Technology School of Computer Science and Engineering Xi'an710048 China National Cheng Kung University Department of Engineering Science Tainan70142 Taiwan

As many countries face the challenges of an aging population and declining birth rates, the demand for labor, particularly for assisting the elderly, is increasing. Traditional robots, being standardized products, require engineers to adjust parameters to fit the unique lifestyle of each household, which is time-consuming and inconvenient. However, with the rapid development of artificial intelligence and consumer electronics, there is a growing need for home robots with smarter interfaces to achieve the goal of intelligent living. This paper proposes a home robot based on a trustworthy AI framework, integrated with large language models (LLM). These LLM can perform natural language processing and object recognition, allowing users to control the robot's operations through natural language commands. This innovation further advances consumer electronics. The robot's arm can remember these actions and operate according to instructions. Additionally, the robot arm is equipped with monitoring functions, capable of overseeing the operation of other robots and using cameras to detect errors. This development is significant in the field of consumer electronics. The robot also uses Long Short-Term Memory (LSTM) networks to predict the motion paths of the robotic arm, ensuring smooth and efficient operation. This integration of AI and robotics aims to enhance the adaptability and functionality of home robots, making them more suitable for the diverse needs of modern households, improving the quality of life for the elderly, and driving innovation in the consumer electronics field. © 1975-2011 IEEE.

关键词： Robotic arms

来源：评论

学校读者我要写书评

暂无评论

An Efficient Approach for EEG Seizure Detection using CNN with Feature Extraction 7

An Efficient Approach for EEG Seizure Detection using CNN wi...

引用

7th International Conference on Circuit Power and Computing Technologies, ICCPCT 2024

作者： Joseph, Liya Izudheen, Sminu Narayanana, Uma Nair, Abhinav R Ashok Gopal, K.A. Rajagiri School of Engineering & Technology Department of Computer Science and Engineering Ernakulam India CVV Institute of Science and Technology Department of Computer Science and Engineering Ernakulam India

ISBN: (纸本)9798350372816

Epileptic seizures are a significant agony for those who suffer from them. Epileptic studies primarily focus on understanding the abnormal behavior of brain signals. Detecting seizures in EEG signals manually is a very chaotic and challenging task. Therefore, we propose an automatic seizure detection system that drastically reduces the manual effort required for seizure detection while providing accurate results. The proposed method involves extracting twelve statistical features from segmented data and adjusting labels to facilitate binary classification using a Convolutional Neural Network (CNN). The dataset is normalized and balanced using SMOTE to address class imbalance. Evaluation metrics such as accuracy, precision and recall are computed on the test set with results visualized through accuracy and loss curves. User input allows for predictions on specific test data, determining the presence of seizures based on model predictions. This structured approach ensures reliable seizure detection through the use of advanced deep learning techniques and comprehensive model evaluation. Furthermore, a comparative study was conducted to assess the effectiveness of the proposed model against conventional machine learning algorithms such as Logistic Regression, Support Vector Machine (SVM), Gradient Boosting and Decision Tree. Our analysis reveals that CNN consistently demonstrates superior performance compared to these methods. © 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

Behavior Based Group Recommendation from Social Media Dataset by Using Deep Learning and Topic Modeling

引用

SN computer science 2024年第6期5卷 1-17页

作者： Mukta, Md. Saddam Hossain Ahmed, Jubaer Raiaan, Mohaimenul Azam Khan Fahad, Nur Mohammad Islam, Muhammad Nazrul Imtiaz, Nafiz Islam, Md. Adnanul Ali, Mohammed Eunus Azam, Sami LUT School of Engineering Sciences Lappeenranta-Lahti University of Technology Department of Computer Science and Engineering United International University Department of Computer Science and Engineering Military Institute of Science and Technology Department of HCC Action Lab Monash University Department of Computer Science and Engineering Bangladesh University of Engineering and Technology Faculty of Science and Technology Charles Darwin University

In this digital era, users frequently share their thoughts, preferences, and ideas through social media, which reflect their Basic Human Values. Basic Human Values (aka values) are the fundamental aspects of human behavior, which define what we consider important, and worth having and pursue them. Existing studies identify the values of individuals from different social network usages such as Facebook and Reddit. However, discovering the similarity (or diversity) of value priorities among the members in a group is important since we can reveal many interesting insights such as finding a set of target customers, identifying the chain of misdeed groups, searching for similar acquaintances in workplaces, etc. In this paper, a graph dataset is compiled using the strongest correlation among the features and then we apply a graph clustering technique to identify a suitable hedonist group (i.e., one dimension of values) for users’ recommendations. Then, we also propose a behavior based (i.e., value) group recommendation technique by analyzing users’ contextual and psychological attributes. Finally, we validate those group members in real life by introducing two hypotheses. In particular, we analyze the tweets of a total of 1140 users collected from Twitter. We obtain a substantial intra-cluster correlation coefficient (ICC) and silhouette clustering coefficient (SCC) scores of 65% and 76%, respectively, among the members in our discovered group. © The Author(s) 2024.

关键词： Graph neural network (GNN) Group recommendation Psychological attributes Social media Topic modeling Twitter Values

来源：评论

学校读者我要写书评

暂无评论

A Vulnerability Detection Method for Smart Contracts Based on Dynamic Meta Optimizer

引用

IEEE Internet of Things Journal 2025年第13期12卷 24904-24915页

作者： He, Daojing Gong, Wei Chan, Sammy Harbin Institute of Technology School of Computer Science and Technology Shenzhen China Jiangxi University of Science and Technology School of Information Engineering Jiangxi Ganzhou China City University of Hong Kong Department of Electrical Engineering Hong Kong Hong Kong

With the increasingly complex blockchain technology environment and emerging security threats, the detection and prevention of vulnerabilities in blockchain smart contracts have become crucial for ensuring the healthy development of blockchain technology and avoiding substantial economic losses. Recently developed vulnerability detection methods for smart contracts suffer from the drawbacks of insufficient feature extraction and inadequate multitask detection. This paper proposes a multifaceted learning model, DLR, based on a dual-loop architecture of meta learning to address these issues by adopting model-agnostic meta-learning techniques. This model employs a syntax analyzer for targeted feature extraction, with graph information used as a supporting tool during the initial stage of feature engineering. Combined with an improved optimizer algorithm in the inner loop, the model can effectively learn and adjust based on the specific requirements of each task. In the outer loop, the model achieves efficient learning rate adjustment for multi-task learning by integrating the adaptive learning rate of the Adam optimizer with a global adjustment using simulated annealing strategy, thereby enhancing performance across diverse tasks. Experimental results demonstrate significant improvements in detection accuracy over state-of-the-art methods for three types of vulnerabilities, with our method achieving detection accuracies of 94.40%, 93.36%, and 94.33% for reentrancy, timestamp dependence, and integer overflow vulnerabilities, respectively. © 2014 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Resolution-aware beam scanning for joint detection and communication in ISAC systems

引用

science China(Information sciences) 2025年第5期68卷 250-266页

作者： Yiming XU Dongfang XU Zhanyuan XIE Zheng JIANG Shenghui SONG Derrick Wing Kwan NG Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology China Telecom Research Institute School of Electrical Engineering and Telecommunications University of New South Wales

Beam scanning for joint detection and communication in integrated sensing and communication(ISAC) systems plays a critical role in continuous monitoring and rapid adaptation to dynamic environments. However, the design of sequential scanning beams for target detection with the required sensing resolution has not been tackled in the *** bridge this gap, this paper introduces a resolution-aware beam scanning design. In particular, the transmit information beamformer, the covariance matrix of the dedicated radar signal, and the receive beamformer are jointly optimized to maximize the average sum rate of the system while satisfying the sensing resolution and detection probability requirements.A block coordinate descent(BCD)-based optimization framework is developed to address the non-convex design problem. By exploiting successive convex approximation(SCA), S-procedure, and semidefinite relaxation(SDR), the proposed algorithm is guaranteed to converge to a stationary solution with polynomial time complexity. Simulation results show that the proposed design can efficiently handle the stringent detection requirement and outperform existing antenna-activation-based methods in the literature by exploiting the full degrees of freedom(DoFs) brought by all antennas.

关键词： integrated sensing and communication (ISAC) detection beam scanning resolution beamwidth

来源：评论

学校读者我要写书评

暂无评论

A Novel Access Point Deployment Framework for mmWave Cell-Free Massive MIMO Networks

引用

IEEE Transactions on Wireless Communications 2025年第6期24卷 4581-4597页

作者： Topal, Ozan Alp Demir, Ozlem Tugfe Bjornson, Emil Cavdar, Cicek School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Stockholm Sweden Department of Electrical and Electronics Engineering TOBB University of Economics and Technology Ankara Turkey

Millimeter-wave network deployment is an essential and ongoing problem due to the limited coverage and expensive network infrastructure. In this work, we solve a joint network deployment and resource allocation optimization problem for a mmWave cell-free massive MIMO network considering indoor environments. The objective is to minimize the number of deployed access points (APs) for a given environment, bandwidth, AP cooperation, and precoding scheme while guaranteeing the rate requirements of the user equipments (UEs). Considering coherent joint transmission (C-JT) and non-coherent joint transmission (NC-JT), we solve the problem of AP placement, UE-AP association, and power allocation among the UEs and resource blocks jointly. For numerical analysis, we model a mid-sized airplane cabin in ray-tracing as an exemplary case for IDS. Results demonstrate that a minimum data rate of 1 Gbps can be guaranteed with less than 10 APs with C-JT. From a holistic network design perspective, we analyze the trade-off between the required fronthaul capacity and the processing capacity per AP, under different network functional split options. We observe an above 600 Gbps fronthaul rate requirement, once all network operations are centralized, which can be reduced to 200 Gbps under physical layer functional splits. 2002-2012 IEEE.

关键词： Integer programming

来源：评论

学校读者我要写书评

暂无评论

Exploratory Data Analysis using Autoviz for Machine Learning Classification Problem

Exploratory Data Analysis using Autoviz for Machine Learning...

引用

2024 International Conference on Emerging Innovations and Advanced Computing, INNOCOMP 2024

作者： Guru Prasad, M.S. Praveen Gujjar, J. Kotiyal, Arnav Prasanna Kumar, H.R. Devadas, Raghavendra M Jahan, Afsar Department of Computer Science and Engg. Dehradun India Faculty of Management Studies Bengaluru India Visvesvaraya Technological University PESITM Dept. of Information Science & Engg. Shivamogga India Gitam School of Technology Department of Computer Science and Engineering Gitam University Bengaluru India Department of Computer Application Dehradun India

ISBN: (纸本)9798350376470

As the field of machine learning continues to advance, the importance of effective exploratory data analysis (EDA) cannot be overstated, especially in the context of classification problems. This research paper introduces an innovative and automated approach to exploratory data analysis using Autoviz, a powerful Python library designed to streamline the visualization process for machine learning datasets. The study aims to demonstrate the efficacy of Autoviz in meaningful insights, facilitating feature understanding, and supporting data preprocessing for classification tasks. Furthermore, the paper discusses how Autoviz assists in feature selection by highlighting the relevance and importance of each feature about the target variable. This feature-centric analysis contributes to optimizing machine learning models by guiding the selection of relevant input features, thereby improving model accuracy and interpretability. Autoviz is a Python library designed to automate the exploratory data analysis (EDA) process for machine learning tasks. It is specifically tailored to generate a variety of visualizations that aid in understanding the structure and characteristics of a dataset. The results indicate that Autoviz significantly reduces the time and effort required for exploratory data analysis while maintaining the quality and depth of insights obtained. The differences between Autoviz and Pycaret is discussed in this paper. It has been observed that AutoViz automates the process of exploring and analyzing data, while PyCaret is designed for the comparison, construction, and tuning of machine learning models. © 2024 IEEE.

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：