检索结果-内蒙古大学图书馆

5th International Conference on Circuits, Control, Communication and Computing, I4C 2024

作者： Priyadarshini, V. Srinivasulu Reddy, U. Reddy Chirra, Venkata Rami Mrudula, M. Suneetha, M. National Institute of Technology Dept. of Computer Science & Engineering Andhra Pradesh India Machine Learning & Data Analytics Lab National Institute of Technology Dept. of Computer Applications Tiruchirappalli India School of Computer Science & Engineering VIT-AP University Amaravathi India Malla Reddy College of Engineering Department of CSE Telangana India Independent Researcher Amaravati India

ISBN: (纸本)9798331528539

Emotions play a crucial role in decision-making, moral judgments, and other cognitive processes. The goal of this work is to identify people's facial emotions from face images. Facial expression recognition (FER), a highly diverse field with many commonalities between classes, is a challenging problem. The capabilities of current FER systems are limited in this regard. In this work, a multi-block Deep Convolutional Neural Networks (DCNN) model was proposed to identify facial expressions. Five blocks with different computational components were defined in a multi-block DCNN to extract the unique characteristics from input images. The second model (DCNN-SVM) is developed with an ensemble of SVM classifier to improve predictability and stability. The third model (DCNN-MV) is based on the majority voting with 4 classifiers SVM, KNN, Random Forest, and Decision Tree. The CK+ dataset was utilized and expanded via image data augmentation to enhance model performance and generalization. By adjusting hyperparameters, the suggested DCNN model's accuracy was investigated. The DCNN model reports an accuracy score of 97.71%. Both DCNN-SVM and DCNN-MV models achieved 97.79% accuracy. On the CK+ dataset, the proposed models demonstrated good accuracy in identifying facial expressions. When compared to existing models, the proposed model outperformed existing models in terms of accuracy. © 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

EUDRL: Explainable Uncertainty-Based Deep Reinforcement Learning for Portfolio Management 15

EUDRL: Explainable Uncertainty-Based Deep Reinforcement Lear...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Lina, Jia Banda, Shahab S. Paib, Hao-Ting Rawal, Bharat S. International Graduate School of Artificial Intelligence Yunlin County Douliu City Taiwan Bachelor Program of Big Data Applications National Pingtung University Pingtung Taiwan Department of Computer Science and Digital Technologies University GgramblingLA United States

ISBN: (纸本)9798350370249

Supervised deep learning (SDL) has shown remarkable success in various financial applications, such as stock prediction and fraud detection. However, SDL's reliance on class labels renders it unsuitable for portfolio management (PM) tasks, where such labels are often unavailable. To address this limitation, we propose a novel two-level architecture based on deep reinforcement learning (DRL) for PM, which does not require class labels. Our approach comprises several local agents that provide trading decisions and uncertainty assessments for individual stocks, and a global agent that makes portfolio management decisions based on the outputs of the local agents. Additionally, we incorporate the concept of explainable AI (XAI) into our framework using the SHAP (Shapley additive explanations) method, enhancing the transparency and interpretability of the global agent's decisions. Our experimental results demonstrate that the proposed architecture consistently yields profitable outcomes in the market. ©2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Key Clues Guided Video Character Social Relationship Recognition Enhanced by LLM

Key Clues Guided Video Character Social Relationship Recogni...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Dong, Wenlong Zhu, Qing Mao, Qirong School of Computer Science and Communication Engineering Jiangsu University China Jiangsu Engineering Research Center of Big Data Ubiquitous Perception and Intelligent Agriculture Applications Provincial Key Laboratory of Computational Intelligence and New Technologies in Low-Altitude Digital Agriculture Zhenjiang China

ISBN: (纸本)9798350368741

Video Character Social Relationship Recognition (VCSRR) requires a comprehensive consideration about spatio-temporal and multi-modal clues in videos. Most existing methods mainly focus on integrating multi-modal clues and modeling interactions among characters. However, they fail to discover key clues in the complex video data or fully understand the clues related to social relationships. In this article, we propose a novel Large Language Model Enhanced Key Clues Selection (LEKCS) framework to address the aforementioned issues. The core of LE-KCS is to mine multi-scale key clues from the perspectives of time, space and multi-modality, then transfer the knowledge about social relationships of the Large Language Model to VCSRR for understanding the selected clues. We evaluated LEKCS on the MovieGraphs dataset and the experimental results indicate that our proposed LE-KCS achieves state-of-the-art performance. © 2025 IEEE.

关键词： large language model Social relationship recognition video understanding

来源：评论

学校读者我要写书评

暂无评论

Does the Fairness of Your Pre-Training Hold Up? Examining the Influence of Pre-Training Techniques on Skin Tone Bias in Skin Lesion Classification

Does the Fairness of Your Pre-Training Hold Up? Examining th...

引用

IEEE Winter applications and computer Vision Workshops (WACVW)

作者： Pratinav Seth Abhilash K. Pai Department of Data Science & Computer Applications Manipal Institute of Technology Manipal Academy of Higher Education Manipal India

Deep Neural Networks (DNNs) have found widespread application in various domains, but the challenge of addressing Algorithmic bias and ensuring fairness in their decision-making processes has emerged as a critical concern, particularly in mission-critical contexts. One of the main reasons for this concern is the inadequate representation of certain groups in the available datasets used for training. Pre-Training is a powerful technique for training DNNs, but it can be affected by pre-existing biases in the dataset. These biases can be transferred to the DNN during Pre-Training, leading to the DNNs making biased decisions, even when trained on unbiased datasets. This study investigates the impact on the fairness of popular Pre-Training methods, such as Masked Image Modeling (MAE, SimMIM) and Self-Supervised Learning (BYOL, MoCo, SimCLR, VICRegL), when used on skin lesion classification datasets with underrepresented demographic groups. The study compares the performance of pre-trained models to supervised learning backbones on two skin lesion datasets (ISIC-2019 and Fitzpatrickl17k) with different skin tone distributions. The findings of this study reveal that Pre-Training improves performance but has a trade-offwith fairness, which can be a potential danger associated with the model when applied in the real world. This study is one of the first to investigate how Self-Supervised Learning and Masked Image Modeling Pre-Training methods affect fairness in both in-distribution and out-of-distribution scenarios. Code is available at https://***/ptnv-s/PretrainingImpactOnSkinBias.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Tabular Learning: Encoding for Entity and Context Embeddings

arXiv

引用

arXiv 2024年

作者： Reusser, Fredy Institute for Data Applications and Security School of Engineering and Computer Science Bern University of Applied Sciences Biel Switzerland

Examining the effect of different encoding techniques on entity and context embeddings, the goal of this work is to challenge commonly used Ordinal encoding for tabular learning. Applying different preprocessing methods and network architectures over several datasets resulted in a benchmark on how the encoders influence the learning outcome of the networks. By keeping the test, validation and training data consistent, results have shown that ordinal encoding is not the most suited encoder for categorical data in terms of preprocessing the data and thereafter, classifying the target variable correctly. A better outcome was achieved, encoding the features based on string similarities by computing a similarity matrix as input for the network. This is the case for both, entity and context embeddings, where the transformer architecture showed improved performance for Ordinal and Similarity encoding with regard to multi-label classification tasks. © 2024, CC BY-SA.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Cognitive Detection of Anomalies in Autonomous In-Vehicle Network Communication

Cognitive Detection of Anomalies in Autonomous In-Vehicle Ne...

引用

Advancements in Power, Communication and Intelligent Systems (APCI), International Conference on

作者： Suzen Firasta Yash Rahul Srivastava Vidya Rao Department of Data Science and Computer Applications Manipal Institute of Technology Manipal Academy of Higher Education Manipal India

ISBN: (数字)9798350363289

ISBN: (纸本)9798350363296

The research addresses the critical issue of detecting anomalies in autonomous in-vehicle network communication, focusing on the vulnerability of infotainment systems to denial-of-service attacks. The aim is to explore and compare traditional Machine Learning (ML) and advanced Deep Learning (DL) models for the detection of DoS attacks within Controller Area Network (CAN) communications. By evaluating the performance of various ML and DL algorithms, including Decision Tree, Logistic Regression, RandomForest, Feed Forward Neural Networks, and LSTM architectures, the study aims to enhance intrusion detection capabilities in vehicular communication systems. The significance of the research lies in identifying the advantages of DL methods in capturing complex patterns and temporal dependencies within the CAN network, thus contributing to fortifying the security of automotive networks against evolving threats. The results demonstrate the efficacy of the proposed models in detecting and mitigating DoS attacks, with DL methods showing promise in automatic learning and adaptability to intricate network structures. This study provides valuable insights for future research endeavors aimed at enhancing the cybersecurity of autonomous vehicles.

关键词： Deep learning Adaptation models Soft sensors Prevention and mitigation Intrusion detection Training data Denial-of-service attack

来源：评论

学校读者我要写书评

暂无评论

Big data Analytics for Quality Control in The Pharmaceutical Industry 10

Big Data Analytics for Quality Control in The Pharmaceutical...

引用

10th International Conference on Advanced Computing and Communication Systems, ICACCS 2024

作者： Malathi, P. Jansi, S. Prabhakaran, M.V. Senbagam, K. Sujatha, M.P. Sadaieswaran, R. Saveetha College of Liberal Arts and Sciences Saveetha Institute of Medical and Technical Sciences Department of Data Analytics Tamil Nadu Chennai India Acharya college of Engineering and Technology Department of Computer science and Engineering Achariyapuram Puducherry India Dhanalakshmi Srinivasan College of Engineering and Technology Department of Computer Science and Engineering Tamil Nadu Mamallapuram India Agni College of Technology Department of Computer Science and Engineering Chennai Thalambur India Tagore College of arts and science Department of Computer Applications Chennai Chromepet India

ISBN: (纸本)9798350384369

In order to ensure the safety and manufacturing efficiency of pharmaceuticals, rigorous quality control measures are necessary. However, existing quality control systems often struggle with industrial process complexity and regulatory requirements. The proposed study highlights the need for a revolutionary strategy to overcome the limitations of existing systems. Quality control systems in existing systems have limitations such as real-time monitoring, insufficient data analysis for pattern identification, and limited predictive capabilities. These limitations have an impact on the industry's ability to recognize and manage quality concerns, posing a danger to medicine manufacturing. For addressing these issues, a novel approach to maintain quality control in the pharmaceutical sector using big data analytics is proposed. Advanced analytics are used for real-time monitoring, data analysis and predictive modeling in the proposed system. The results and analysis through extensive comparisons shows proposed big data analytics system achieves 95% fault detection, 40% downtime reduction, and 35% waste reduction in pharmaceutical manufacturing. The proposed system's ability to handle massive datasets allows it to identify complicated patterns and abnormalities, enabling proactive quality control. The implementation exceeds industry standards, addressing data privacy and security concerns. © 2024 IEEE.

关键词： data quality

来源：评论

学校读者我要写书评

暂无评论

Classification and Discovery of Brain Tumor Using Ensemble Deep Learning 2

Classification and Discovery of Brain Tumor Using Ensemble D...

引用

2nd IEEE International Conference on Networks, Multimedia and Information Technology, NMITCON 2024

作者： Srilatha, Kasanagottu Sravani, Chekka Vasudevarao, Mikkilineni Kiran, Ajmeera Chitra, D. Reddy, Pundru Chandra Shaker Geethanjali College of Engineering and Technology Department of Data Science Hyderabad India Marri Laxman Reddy Institute of Technology and Management Computer Science and Engineering Telangana Hyderabad India Dayananda Sagar Business Academy Department Of Computer Applications Karnataka India Mlr Institute of Technology Computer Science and Engineering Dundigal Telangana Hyderabad India Panimalar Engineering College Master of Business Administration Tamil Nadu Chennai India Sr University School of Computer Science and Artificial Intelligence Telangana Warangal India

ISBN: (纸本)9798350372892

As healthcare IT develops at a dizzying rate, the age of big data is drawing near. Tumor disease prediction, monitoring, diagnosis, and treatment are all profoundly affected by data mining and analysis. Brain-tumors(BT) are considered the most terrible and deadly disease due to their aggressive nature, short survival rate, and diverse spectrum of symptoms. Inadequate medical therapy for patients with misdiagnosed brain tumors lowers their survival prospects. The ability to differentiate amid abnormal &normal tissues makes BT detection extremely difficult. An accurate diagnosis allows the patient to undergo effective therapy and has a better chance of long-term survival. Because of the abnormal pattern of lesion distribution, there are still convinced difficulties in diagnosing BTs, even though there has been a lot of research on the subject. Because tiny areas may seem healthy, it might be challenging to locate a zone with few lesions. Classification accuracy is negatively impacted, and selecting and extracting useful characteristics becomes more difficult. An important part is the automated categorization of precancerous BTs utilizing deep &machine learning strategies. This study presents a Convolutional-Neural-Network-Long-Short-Term-Memory (CNN-LSTM) model, a hybrid DL-technique, for the purpose of MRI-based brain tumor classification and prediction. An MRI brain imaging collection is utilized for our experiments. An effective preprocessing of the data is the first step in using a CNN to extract major characteristics from images. With a 99.3% significant classification accuracy, 99.2% precision, 99.1% recall, and 99.1% F1-measure, the suggested model successfully predicts the brain tumor. © 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

Exploring Sensor-Based Smart Farming Technologies in the Internet of Things (IoT) 1

Exploring Sensor-Based Smart Farming Technologies in the Int...

引用

1st IEEE International Conference on Computing and data science, ICCDS 2024

作者： Palarimath, Suresh Maran, Pyingkodi Thenmozhi, K. Balakumar, C. Sujatha, T. Wilfred Blessing, N.R. College of Computing and Information Sciences University of Technology and Applied Sciences Information Technology Department Salalah Oman Kongu Engineering College Department of Computer Applications Erode India Department of Compute Science Bengaluru India Karpagam Academy of Higher Education Department of Computer Science Coimbatore India Sri Krishna College of Engineering and Technology Department of Artificial Intelligence and Data Science Coimbatore India

ISBN: (纸本)9798350365337

The Internet of Things (loT) is a new technology trend that is being used in almost every area of human life. IoT is used almost every aspect of people's lives. Significantly, with a projected increase in the world's population to 9.7 billion by 2050, agricultural output would need to increase at an even more rapid rate to fulfill the requirement. Modern tools, notably the Internet of Things, make this a reality. The IoT makes it possible for farms to function without human labor. It has several potential applications in agriculture, including large- scale farming, greenhouse farming and management. The sensors serve as the most crucial component of the loT. Sensing devices are mostly used for the purpose of learning about the soil and its surroundings. The sensor has several applications in agriculture, including but not limited to NPK (nitrogen, phosphorus, and potassium) measurement, disease detection, and soil moisture analysis. This study discusses how IoT applications contribute to efficient farming practices. It shows how the IoT can be applied to agriculture and exhibits the many sensors, applications, problems, strengths, and shortcomings that underpin this field. © 2024 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Efficient mode decision in HEVC

Efficient mode decision in HEVC

引用

2022 International Conference on Artificial Intelligence and data Engineering, AIDE 2022

作者： Bijur, Gururaj Ramakrishna, M. Karunakar, A.K. Manipal Institute of Technology Department of Computer Science & Engineering MAHE Manipal Karnataka Udupi India Manipal Institute of Technology Department of Data Science & Computer Applications MAHE Manipal Karnataka Udupi India

ISBN: (纸本)9781665493048

HEVC is the new coding standard aiming to provide better coding efficiency compared to its previous standard. Meantime this increases the computational complexity. In HEVC, there are 35 intra prediction modes, where as in H.264/A VC there were only 9 intra prediction modes. In this paper, we proposed a fast rough mode decision (RMD) algorithm to decrease the computational complexity using up and left reference pixels. The three-stage search method is implemented in RMD algorithm. The results obtained when compared with HM16.18, shows that the proposed algorithm can save 12.78% to 28.86% of the encoding time with acceptable loss in video quality. © 2022 IEEE.

关键词： Computational complexity

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：