检索结果-内蒙古大学图书馆

41st International Conference on machine learning, ICML 2024

作者： Steinmann, David Stammer, Wolfgang Friedrich, Felix Kersting, Kristian Artificial Intelligence and Machine Learning Group TU Darmstadt Germany Darmstadt Germany Centre for Cognitive Science TU Darmstadt Germany Germany

While deep learning models often lack interpretability, concept bottleneck models (CBMs) provide inherent explanations via their concept representations. Moreover, they allow users to perform interventional interactions on these concepts by updating the concept values and thus correcting the predictive output of the model. Up to this point, these interventions were typically applied to the model just once and then discarded. To rectify this, we present concept bottleneck memory models (CB2Ms), which keep a memory of past interventions. Specifically, CB2Ms leverage a two-fold memory to generalize interventions to appropriate novel situations, enabling the model to identify errors and reapply previous interventions. This way, a CB2M learns to automatically improve model performance from a few initially obtained interventions. If no prior human interventions are available, a CB2M can detect potential mistakes of the CBM bottleneck and request targeted interventions. Our experimental evaluations on challenging scenarios like handling distribution shifts and confounded data demonstrate that CB2Ms are able to successfully generalize interventions to unseen data and can indeed identify wrongly inferred concepts. Hence, CB2Ms are a valuable tool for users to provide interactive feedback on CBMs, by guiding a user's interaction and requiring fewer interventions. Copyright 2024 by the author(s)

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

Comparative Assessment of machine learning Models for Predicting Glucose Intolerance Risk

引用

SN computer science 2024年第7期5卷 894页

作者： Kumar, B. P. Pradeep Manoj, H.M. Department of Computer Science and Design Atria Institute of Technology Affiliated to VTU Karnataka Bangalore India Department of Artificial Intelligence and Machine Learning BMS Institute of Technology and Management Affiliated to VTU Karnataka Bangalore India

This study focuses on predictive analytics and healthcare, specifically the prediction of Glucose Intolerance, a chronic metabolic disease with significant global health implications. The research aims to develop efficient tools for risk assessment and early identification, offering medical practitioners reliable instruments for identifying high-risk patients and implementing preventive measures in a timely manner. A significant challenge in managing Glucose Intolerance is the absence of effective and precise prediction models. Traditional risk assessment techniques often fail to capture the multifaceted nature of Glucose Intolerance development, leading to delayed treatments and suboptimal patient outcomes. To address this issue, we conducted a comprehensive study utilizing various machine learning algorithms, including Decision Trees, Random Forest, Gradient Boosting, CatBoost, K-Nearest Neighbors (KNN), Support Vector Classifier (SVC), Logistic Regression, and a Voting Classifier, to predict Glucose Intolerance. The primary objective was to identify the most effective combination of features and models for accurate predictions. Key factors considered included patient demographics, lifestyle characteristics, medical history, and genetic susceptibility, which were used to build robust and personalized prediction models. We conducted a comparative analysis of the machine learning models' performance based on cross-validated accuracy with test-train splits and folds: 0.2, 6;0.05, 4;0.1, 2;0.2, 6;and 0.05, 2. The results showed that Random Forest achieved test accuracies of 74%, 76%, 82%, 75%, and 79%;KNN achieved 70%, 72%, 74%, 75%, and 72%;SVC achieved 74%, 72%, 77%, 77%, and 72%;Logistic Regression achieved 73%, 76%, 79%, 73%, and 76%;Gradient Boosting achieved 72%, 79%, 75%, 73%, and 79%;XGBoost achieved 72%, 79%, 77%, 74%, and 79%;CatBoost achieved 74%, 76%, 79%, 75%, and 76%;Decision Tree achieved 60%, 83%, 74%, 70%, and 79%;and Voting Classifier achieved 74%, 7

关键词： Algorithmic approaches Chronic disease Data-driven healthcare Early detection Glucose Intolerance Healthcare analytics machine learning Personalized medicine Prediction Predictive modelling Risk stratification

来源：评论

学校读者我要写书评

暂无评论

Node Coverage and Connectivity Optimization in Heterogeneous Wireless Sensor Networks Using Improved Flower Pollination Algorithm 4

Node Coverage and Connectivity Optimization in Heterogeneous...

引用

4th International Conference on Ubiquitous Computing and Intelligent Information Systems, ICUIS 2024

作者： Mahankali, Saritha Kesavan, R. Kalaiselvan, S.A. Saveetha School of Engineering Simats Department of Computer Science & Engineering Chennai India Vardhaman College of Engineering Department of Computer Science & Engineering Telangana India Rajalakshmi Engineering Department of Artificial Intelligence & Machine Learning Chennai India

ISBN: (纸本)9798331529635

Wireless sensor networks (WSNs) rely on energy-conserving routing protocols due to the limited power and communication capabilities. The LEACH protocol has seen extensive usage despite its homogeneous network foundation, leads inadequate signal coverage. To solve the issue, this research proposes an improved flower pollination algorithm (IFPA) improves flower pollination algorithm (FPA) addresses issues with slow convergence and local optima. This research presents an enhanced clustering technique that may be utilized in heterogeneous sensor networks. Its main feature is the ability to adjust the status of data transmission. Once the observed information intensity above the predefined threshold, the data transmission connection is initiated and data is sent to the sink by the cluster heads, which filter the perceived information. Otherwise, cluster nodes keep track of the data and wait for the next batch of nodes to provide data. The good news is that the simulation results show a network lifespan that is several times longer than the LEACH protocol. The IFPA method stands out because it increases coverage rates while drastically decreasing the cost of deploying networks. As a result, the suggested method might find many uses, helping to lower costs and improve the performance of wireless sensor networks. © 2024 IEEE.

关键词： Sensor nodes

来源：评论

学校读者我要写书评

暂无评论

Graph Convolutional Networks for SEO: A Comprehensive Framework for Healthcare Information Ranking

引用

Journal of The Institution of Engineers (India): Series B 2025年 1-16页

作者： RR, Prianka M, Vedaraj R, Anto Arockia Rosaline Department of Artificial Intelligence and Machine Learning New Horizon College of Engineering Bangalore India Department of Computer Science and Engineering R.M.D. Engineering College RSM Nagar Kavaraipettai Thiruvallur India Department of Computing Technologies School of Computing SRM Institute of Science and Technology Kattankulathur Campus Chennai India

Internet’s remarkable surge, ubiquitous accessibility, and serviceability have increased users’ dependency on web services for fast search and recovery of wide sources of information. Search engine optimization (SEO) has become paramount in healthcare industries, which helps patients enhance and understand their health status based on their records. In the context of healthcare, it is more significant to improve search results from specific keywords related to clinical conditions, treatments, and healthcare services. So, this research work proposes a Graph Convolutional Network (GCN)-based Search Engine Optimization (SEO) algorithm for healthcare applications. The algorithm utilizes two distinct datasets: MIMIC-III Clinical Database and Consumer Health Search Queries to optimize search engine rankings for health related queries. Following data acquisition, data pre-processing is performed for better enrichment of analysis. The preprocessing steps involve data cleaning, data integration, feature engineering, and knowledge graph construction procedures to remove noisy data, integrate medical data with user search behavior, compute significant features, and construct knowledge graphs, correspondingly. The relation between the data entities is examined within constructed graph through link analysis. The pre-processed data including medical knowledge weights, content relevance scores, and user interaction signals are processed further on GCN model with Adam-tuned weights and bias for ranking healthcare data based on relevance score in response to user query using cosine similarity. The search relevance estimation indicators namely recall, precision, f1-score, and normalized discounted cumulative gain (NDCG) are computed to measure search optimization performance. The proposed GCN-SEO approach benchmarked its effectiveness over existing methods in optimizing web searches related to healthcare with a high performance rate of 95.75% accuracy and 48.25 s dwell time. This sho

关键词： Clinical research

来源：评论

学校读者我要写书评

暂无评论

StaR Maps: Unveiling Uncertainty in Geospatial Relations 27

StaR Maps: Unveiling Uncertainty in Geospatial Relations

引用

27th IEEE International Conference on Intelligent Transportation Systems, ITSC 2024

作者： Flade, Benedict Kohaut, Simon Eggert, Julian Dhami, Devendra Singh Kersting, Kristian Honda Research Institute Europe GmbH Offenbach63073 Germany Artificial Intelligence and Machine Learning Lab Department of Computer Science Tu Darmstadt Darmstadt64283 Germany Tu Eindhoven Uncertainty in Artificial Intelligence Group Department of Mathematics and Computer Science MB Eindhoven5600 Netherlands Hessian Ai Centre for Cognitive Science Germany

ISBN: (纸本)9798331505929

The growing complexity of intelligent transportation systems and their applications in public spaces has increased the demand for expressive and versatile knowledge representation. While various mapping efforts have achieved widespread coverage, including detailed annotation of features with semantic labels, it is essential to understand their inherent uncertainties, which are commonly underrepresented by the respective geographic information systems. Hence, it is critical to develop a representation that combines a statistical, probabilistic perspective with the relational nature of geospatial data. Further, such a representation should facilitate an honest view of the data's accuracy and provide an environment for high-level reasoning to obtain novel insights from task-dependent queries. Our work addresses this gap in two ways. First, we present Statistical Relational Maps (StaR Maps) as a representation of uncertain, semantic map data. Second, we demonstrate efficient computation of StaR Maps to scale the approach to wide urban spaces. Through experiments on real-world, crowd-sourced data, we underpin the application and utility of StaR Maps in terms of representing uncertain knowledge and reasoning for complex geospatial information. © 2024 IEEE.

关键词： Knowledge representation

来源：评论

学校读者我要写书评

暂无评论

Automated Unified System for LPG Refill Booking & Lekage Detection: A Pervasive Approach 15

Automated Unified System for LPG Refill Booking & Lekage Det...

引用

15th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2024

作者： Suraksha, P. Aishwarya Kumari, Kavana L. Dept.of Computer Science and Engineering Vidyavardhaka College of Engineering Karnataka Mysuru India Dept.of Artificial Intelligence and Machine learning Vidyavardhaka College of Engineering Karnataka Mysuru India

ISBN: (纸本)9798331300579

Civilization is becoming increasingly aware of natural resources, it results in the use of propane and butane, and hence the use of *** India most people use the LPG *** per current government regulations, intense demand but shortage in production of LPG cylinder, once a new cylinder is booked we need to wait for some days to get it *** no cost-efficient technique is available which identifies when cylinder is about to get *** this paper we designed, developed, a proven Automated Unified System using smart Pervasive sensors, where LPG retailer requests for a new cylinder using Short Message Service (SMS) once it identifies cylinder will be empty in few days and also provides detection of gas leakage for better service to the consumers and reduce the scope of irregularities. © Grenze Scientific Society, 2024.

关键词： Liquefied petroleum gas

来源：评论

学校读者我要写书评

暂无评论

Deep learning Approaches for Strengthening Network Intrusion Detection Systems 7

Deep Learning Approaches for Strengthening Network Intrusion...

引用

7th IEEE Pune Section International Conference, PuneCon 2024

作者： Kaulgud, Sanjeev P Chempavathy, B. Tejeswar, Nama Sai Vigneshwar, Chekka Kesava Kumar, Akuthota Rakesh Eswar, Akula New Horizon College of Engineering Department of Artificial Intelligence and Machine Learning Karnataka Bengaluru India New Horizon College of Engineering Department of Computer Science & Engineering Karnataka Bengaluru India

ISBN: (纸本)9798331527822

Strengthening network security is a must in today's digital landscape. Existing Intrusion Detection Systems commonly make use of deep learning techniques such as Deep Neural Networks to identify malicious & anomaly activities. However, there are serious problems of overfitting and huge demand for computational resources in this system that hinder the efficiency and scalability of this system. It introduces an advanced form of an Intrusion Detection System (IDS) which integrates a hybrid machine learning algorithm with Recurrent Neural Network (RNN) and Bidirectional Long Short-Term Memory (Bi-LSTM) along with other machine learning techniques to improve threat detection. This system would realize the recognition of established patterns as well as new anomalies about attacks by means of enhanced feature engineering and real-time AI-driven threat intelligence. It also addresses scalability for more extensive and complex configurations of the network as well as response mechanisms that automatically neutralize threats quickly. Then there is user behavioral analytics that helps in identifying insider threats. It is aimed at providing an all-embracing security solution that is strong and adaptable, greatly improving overall protection of the network. Hence, the effective execution of this project will thus provide a new benchmark for the effectiveness of IDS in delivering reliable security across different sectors as well as paving the way for further improvements in cybersecurity around the globe. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Efficient Chromosome Abnormal Detection using Yolo-V8 model 3

Efficient Chromosome Abnormal Detection using Yolo-V8 model

引用

3rd International Conference on Intelligent Systems, Advanced Computing, and Communication, ISACC 2025

作者： Umarani, V. Isaac, Augustian Premnath, K. Kannan, G. Sanjay, A. Hari Haran, L.K. Saveetha Engineering College Dept of Computer Science and Engineering TamilNadu Chennai India Saveetha Engineering College Dept of Artificial Intelligence & Machine Learning TamilNadu Chennai India

ISBN: (纸本)9798331523893

Human chromosomes exhibit complex and intricate patterns, making their analysis and data interpretation challenging. Studying chromosomes is crucial for identifying genetic disorders, such as Down syndrome, and various types of cancer. Traditional karyotyping methods depend heavily on manual method, which causes human error and high cost. To address these challenges, a deep learning-based approach is proposed to automate the karyotyping process, improving both accuracy and efficiency while minimizing human involvement and errors. This research leverages advanced deep learning techniques, specifically the YOLOv8 object detection framework, to enhance chromosomal analysis. From experiments, the proposed YOLO V8 model achieves a 95.2% accuracy, surpassing other models such as DCNN, CNN, and VGG-19. This work represents a significant step toward developing a more reliable and efficient diagnostic tool for detecting genetic disorders. © 2025 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Acoustic Scene Classification Using ANN and CNN Techniques

Performance Analysis of Acoustic Scene Classification Using ...

引用

2023 International Conference on Integrated intelligence and Communication Systems, ICIICS 2023

作者： Rakshitha Karthik, K. Shetty, Aishwarya D Sowmya, P. Shettigar, Rashmitha Department of Artificial Intelligence & Machine Learning Nitte India School of Computer Science and Engineering Vellore India Department of Computer Science and Engineering Nitte India /CS Dhaisar Mumbai India

ISBN: (纸本)9798350315455

artificial neural networks have achieved notable improvements in various learning tasks and have also been used to classify environmental sounds. The technique for allowing devices to recognize and interpret the surroundings from a recording is called Acoustic Scene Classification (ASC). Computational Auditory/Scene analysis is analyzing scenes using computational techniques. This research paper's primary goal is to categorize audio recordings based on the scenarios or environments in which they were captured. In most applications, deep learning is one of the most recent trends. In this work, we have concentrated on two deep learning techniques employed to classify audio scenes: Convolution Neural Networks (CNN) and artificial Neural Networks (ANN) to develop the model. ReLU and LeakyReLU activation functions are used to train and validate the model. For evaluating the model performance, K-fold cross-validation was used with K=10 and obtained an accuracy of 75.8% using an artificial Neural Network and a proposed model with 90.6% accuracy using a Convolutional Neural Network. Comparing the results with other baseline models that achieved a performance of 73.3% for MCLNN with Mel-spectrogram, 78.2% for Pyramid-Combined CNN, 78.3% for Envnet-v2, and 81.9% for D-CNN. © 2023 IEEE.

关键词： Acoustic Scene Classification ANN CNN Environmental Sound Recognition

来源：评论

学校读者我要写书评

暂无评论

Enhanced Image Processing for Automobile Accident Detection Using Deep learning 3

Enhanced Image Processing for Automobile Accident Detection ...

引用

3rd International Conference on Smart Technologies and Systems for Next Generation Computing, ICSTSN 2024

作者： Usharani, S. Dhanushkumar, J. Manju Bala, P. Balachandar, A. IFET College of Engineering An Autonomous Institution Department of Artificial Intelligence and Machine Learning Villupuram India IFET College of Engineering An Autonomous Institution Department of Computer Science and Engineering Villupuram India

ISBN: (纸本)9798350391565

Road safety can be creatively increased by utilizing systems for reporting and detecting accidents use the YOLO algorithm. Yolo, which stands for "You Only Look Once,"is a sophisticated object recognition system that can identify and pinpoint objects in live video streams. By identifying accidents and notifying emergency services, the system lowers the reaction times and increases the possibility of saving lives by utilizing the YOLO algorithm. The object detection and alarm systems are the two primary parts of the proposed system. The YOLO method is employed by the object recognition module to look for accidents in live video broadcasts. A module of accident photos was used to teach the system to accurately identify incidents. When an accident is discovered, an alert system is activated. The location of the accident and a brief account of what transpired are communicated to the emergency services by the alert system. This information is communicated to emergency services through a wireless communication network, which expedites response times and increases the likelihood of saving lives. Positive results were obtained by testing the system using the images and an accident module. It was shown that the warning system could react in a couple of seconds and that the YOLO algorithm could identify accidents having a precision of around 94. Highways, busy intersections, and other high-risk areas may have systems in place to increase traffic safety and lower the number of accidents. Developing crash detection and warning systems using deep learning using the YOLO technique is one way to potentially increase road safety. With the use of technology, problems may be precisely identified in real-time video feeds, alerting emergency services and potentially increasing survival rates and reaction times. © 2024 IEEE.

关键词： Image enhancement

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：