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： Ajith Kumar, B. Nayana, B.R. Shashank, P. M. S. Ramaiah University of Applied Sciences Dept. of Computer Science and Engineering Bengaluru India Ace Designers Ltd. Research and Technology Development Group Bengaluru India

ISBN: (数字)9798331528539

ISBN: (纸本)9798331528539

computer Numerical Control (CNC) machines are employed in various industrial manufacturing processes to produce precise components. The CNCs are prone to thermal deviations from the heat generated during machining. Axis motors, rotary bearings, and spindle motors are the main sources of heat. In this context, temperature sensitive areas (TSAs) apply to these locations. The accuracy of the machining process is directly affected by the heat-induced deviation between the tool and the workpiece. The objective is to minimize these deviations between the tool and the workpiece. The most challenging aspect of creating thermal error compensation models is figuring out which TSAs to use. This work proposes employing particle swarm optimization (PSO), a computational intelligence (CI) technique, to choose significant features. First, information about the machine's temperature and thermal deviation is collected. Second, the outliers are eliminated by cleaning the data. Following this, PSO optimizes the error band to identify which TSAs are significant. Advanced machine learning (ML) models like polynomial regression (PR) and random forest (RF) regression are trained utilizing these TSAs as features. The proposed approach is compared with matching models developed with features selected through the brute force (BF) approach. While PSO-RF outperformed BF-RF in providing 16.60% error reduction with different features, PSO-PR and BF-PR provided 47.12% error band reduction with similar features. Based on the findings, it can be concluded that PSO-PR can be used on the machine and that PSO outperforms BF in terms of identifying the ideal subset of influential features. © 2024 IEEE.

关键词： Error compensation

来源：评论

学校读者我要写书评

暂无评论

Deep Analysis of Machine Learning Techniques for the Effective Retrieval of Images Based on Content

Deep Analysis of Machine Learning Techniques for the Effecti...

引用

Disruptive Technologies (ICDT), International Conference on

作者： Ashish Jain Sudeep Varshney Dept. of Computer Science and Engineering SET Sharda University Greater Noida India Dept. of Computer Science and Engineering G L Bajaj Institute of Technology and Management Greater Noida India

Our research aims to investigate how machine learning techniques can be applied for efficient image retrieval based on content. It is a very difficult process to obtain accurate images from vast digital image collections. CBIR has been particularly interested in this area. It has been shown that content- or feature-based approaches are more effective than text-based query techniques. Research has shown that the performance of these methods varies depending on the type and size of the data set analyzed, with some algorithms performing well for specific small data sets and others for larger data sets. This article provides a comprehensive and systematic review of effective methods for obtaining the image. The paper's objective is to examine different ML classifiers that are applied via the researchers and compare their results based on accuracy and performance. These algorithms include SVM, KNN, random forest, BoVW, Naive Bayes, and LDA. This review paper can assist researchers in selecting the best classifier for their research, as well as in improving existing classifiers by combining them with other algorithms.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Deep Learning Models in Driver Behavior Recognition

Performance Analysis of Deep Learning Models in Driver Behav...

引用

Cybernation and Computation (CYBERCOM), International Conference on

作者： Rameshwari Kapoor Rashmi Kuliyal Nilesh Bhanot Piyush Bagla Ruchira Rawat Dept. of Computer Science and Engineering Graphic Era Hill University Dehradun Uttarakhand India Dept. of Computer Science and Engineering Graphic Era (Deemed to be University) University Dehradun Uttarakhand India

ISBN: (数字)9798331542344

ISBN: (纸本)9798331542351

The World Health Organization report states that nearly 1.25 million people lose their lives due to an increasing number of road accidents every year. The majority of accidents are reported due to driver errors, which include distractions like the use of mobile phones and fatigue, which causes drowsiness. This study aims to address these issues by performing a comparative analysis between different deep learning models representing two types of learning mechanisms: traditional deep learning and transfer learning. This study portrays the use of two datasets to accomplish this task. These two datasets are the Driver Drowsiness dataset DDD and the State Farm Distracted Drivers dataset. The purpose of this investigative research is to explore the execution and efficiency differences between three deep learning algorithms, the traditional convolutional neural network, ResNet50, and VGG16, in classifying driver behavior by training these models individually on both datasets. This study also targets developing an alarm system that sounds an alarm based on a threshold by using an ensemble of all the models trained for comparative study and averaging their predictions to classify driver behavior. VGG16 achieved an accuracy of 99.74% on the driver drowsiness dataset, whereas Custom CNN Architecture achieved an accuracy of 98.70 % on the driver distraction dataset.

关键词： Deep learning Training Analytical models Accuracy Road accidents Transfer learning Convolutional neural networks Driver behavior Vehicles Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of a Comprehensive RAG-Driven Dashboard Within the ILO System for data Visualization and Query Support: A Case Study of the University of Petra

Design and Implementation of a Comprehensive RAG-Driven Dash...

引用

Computational Intelligence Approaches and Applications (ICCIAA), International Conference on

作者： Mohammad Abu-Arqoub Abed Alkarim Banna Nuha El-Khalili Mohammad Al-Shaikh Hasan Department of Computer Science University of Petra Amman Jordan Data Science & Artificial Intelligence Dept. University of Petra Amman Jordan Software Engineering Department University of Petra Amman Jordan Computer Science Department Brunel University London London UK

ISBN: (数字)9798331523657

ISBN: (纸本)9798331523664

This article presents a comprehensive study on designing and implementing a dashboard for enhanced data visualization and query support at the University of Petra. The system leverages retrieval-augmented generation (RAG) and large language models (LLMs) to support diverse document types, including curricula, course descriptions, and program outcomes, alongside intended learning outcome (ILO)-related files. Our implementation demonstrates significant improvements in data accessibility and query response times while maintaining high accuracy in information retrieval and visualization. Through extensive evaluation, we show that this innovative approach transforms data management processes in higher education by enabling natural language interactions with educational data systems, building upon established business intelligence frameworks while introducing advanced AI capabilities.

关键词： Large language models Retrieval augmented generation Natural languages data visualization Transforms Educational technology Information retrieval data systems Business intelligence Time factors

来源：评论

学校读者我要写书评

暂无评论

Inductive Program Synthesis via Iterative Forward-Backward Abstract Interpretation

arXiv

引用

arXiv 2023年

作者： Yoon, Yongho Lee, Woosuk Yi, Kwangkeun Seoul National University Dept. of Computer Science & Engineering Korea Republic of Hanyang University Dept. of Computer Science & Engineering Korea Republic of

A key challenge in example-based program synthesis is the gigantic search space of programs. To address this challenge, various work proposed to use abstract interpretation to prune the search space. However, most of existing approaches have focused only on forward abstract interpretation, and thus cannot fully exploit the power of abstract interpretation. In this paper, we propose a novel approach to inductive program synthesis via iterative forward-backward abstract interpretation. The forward abstract interpretation computes possible outputs of a program given inputs, while the backward abstract interpretation computes possible inputs of a program given outputs. By iteratively performing the two abstract interpretations in an alternating fashion, we can effectively determine if any completion of each partial program as a candidate can satisfy the input-output examples. We apply our approach to a standard formulation, syntax-guided synthesis (SyGuS), thereby supporting a wide range of inductive synthesis tasks. We have implemented our approach and evaluated it on a set of benchmarks from the prior work. The experimental results show that our approach significantly outperforms the state-of-the-art approaches thanks to the sophisticated abstract interpretation techniques. © 2023, CC BY.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Enhancing Split Computing and Early Exit Applications through Predefined Sparsity

arXiv

引用

arXiv 2024年

作者： Capogrosso, Luigi Fraccaroli, Enrico Petrozziello, Giulio Setti, Francesco Chakraborty, Samarjit Fummi, Franco Cristani, Marco Dept. of Engineering for Innovation Medicine University of Verona Italy Dept. of Computer Science University of Verona Italy Dept. of Computer Science University of North Carolina Chapel Hill United States

In the past decade, Deep Neural Networks (DNNs) achieved state-of-the-art performance in a broad range of problems, spanning from object classification and action recognition to smart building and healthcare. The flexibility that makes DNNs such a pervasive technology comes at a price: the computational requirements preclude their deployment on most of the resource-constrained edge devices available today to solve real-time and real-world tasks. This paper introduces a novel approach to address this challenge by combining the concept of predefined sparsity with Split Computing (SC) and Early Exit (EE). In particular, SC aims at splitting a DNN with a part of it deployed on an edge device and the rest on a remote server. Instead, EE allows the system to stop using the remote server and rely solely on the edge device’s computation if the answer is already good enough. Specifically, how to apply such a predefined sparsity to a SC and EE paradigm has never been studied. This paper studies this problem and shows how predefined sparsity significantly reduces the computational, storage, and energy burdens during the training and inference phases, regardless of the hardware platform. This makes it a valuable approach for enhancing the performance of SC and EE applications. Experimental results showcase reductions exceeding 4 × in storage and computational complexity without compromising performance. The source code is available at https://***/intelligolabs/sparsity_sc_ee. © 2024, CC BY.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

On Rejecting Low Quality Images to Improve Deep Smartphone Wound Assessment 22

On Rejecting Low Quality Images to Improve Deep Smartphone W...

引用

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Ditthapron, Apiwat Agu, Emmanuel O. Pedersen, Peder Lindsay, Clifford Tulu, Bengisu Strong, Diane Computer Science Dept. WorcesterMA United States Electrical and Computer Engineering Dept. WorcesterMA United States Univ. Massachusetts Medical School Department of Radiology WorcesterMA United States WorcesterMA United States

ISBN: (纸本)9798350345346

Automatic assessment of wounds using smartphones can facilitate the tracking of healing progress outside the clinic. Prior work proposed SS-PMG-EfficientNets, a state-of-the-art, multi-scale, convolutional neural network architecture for accurate wound assessment. However, the model was previously only trained and evaluated on high-quality wound images captured in controlled environments. This study systematically evaluated the effects of low-quality images on SS-PMG-EfficientNets, finding that the model's performance and consistency are reduced when input wound images are blurry or captured in adverse lighting (too bright or too dark), two factors previously identified as the most prevalent issues in crowdsourced images. Specifically, SS-PMG-EfficientNet was first trained on three high-quality wound datasets containing images labeled with Photographic Wound Assessment Tool (PWAT) wound healing scores, stages of pressure injury wounds, and wound types, respectively. The model was then evaluated on a test wound dataset that was augmented with real-world image luminance, blur, and image compression effects and found that wound assessment accuracy decreased by up to 27%. To improve model performance, we propose a method for instant image quality assessment on the smartphone, enabling low-quality wound images to be rejected immediately after image capture so that the nurse or patient can be prompted to recapture the image. Our evaluation of the proposed image quality assessment and rejection method showed that it increases average accuracy from 73 % to 87%. © 2023 IEEE.

关键词： Chronic wounds Image quality assessment Smart health

来源：评论

学校读者我要写书评

暂无评论

Enhancing Dietary Monitoring using Deep Learning: Food Recognition and Calorie Estimation

Enhancing Dietary Monitoring using Deep Learning: Food Recog...

引用

International Conference on Computing and Networking Technology (ICCNT)

作者： M. Suneela Mandadi Sahithi Nuthalapati Vamsi Nunsavathu Balaji Naik Narra Sujitha Chowdary Dept. of Computer Science & Engineering Vignan’s Lara Institute of Technology & Science College Guntur India

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

The project “Enhancing Dietary Monitoring Using Deep Learning:Food Recognition And Calorie Estimation” introduces an innovative method aimed at improving dietary tracking and fostering healthier eating habits. By employing cutting-edge deep learning techniques, its core objective is to precisely identify various food items and rapidly estimate their caloric content, offering users advanced and personalized monitoring capabilities. Utilizing the Python programming language alongside the MobileNet architecture model, the project underwent rigorous training and evaluation using the expansive Food 101 dataset, comprising 37,046 food images distributed across 101 distinct classes. Notably, the model demonstrated exceptional performance, achieving a training accuracy of $\mathbf{9 7 . 0 2 \%}$ and a validation accuracy of $\mathbf{9 8 . 1 7 \%}$, highlighting the efficacy of this approach in accurately categorizing a wide array of food items. This system provides users with several crucial functionalities. Furthermore, it furnishes users with essential insights by estimating the caloric content of recognized foods, facilitating effective monitoring of dietary intake. The intelligent diet monitoring capabilities enabled by Food Recognition and Calorie Estimation empower users to make informed choices regarding their dietary preferences. Through the continuous tracking and analysis of their daily food consumption, users can glean valuable insights into their nutritional habits, set personalized goals, and make necessary adjustments to achieve a well-balanced and healthy diet. Enhancing Dietary Monitoring Using Deep Learning:Food Recognition And Calorie Estimation stands as a prominent illustration of successful implementation of deep learning techniques, particularly with the utilization of the MobileNet architecture, for food recognition and calorie estimation. With its remarkable accuracy, real-time processing capabilities, and intelligent monitoring features, this projec

关键词： Deep learning Training Accuracy Scalability Estimation computer architecture Real-time systems Object recognition Monitoring Python

来源：评论

学校读者我要写书评

暂无评论

Bayesian Belief Network Analysis for SPAD System in Railways 4

Bayesian Belief Network Analysis for SPAD System in Railways

引用

4th International Conference on Multimedia Processing, Communication and Information Technology, MPCIT 2024

作者： Krishnamurthy, Manjunath Sargur Rahul, Bommisetty Bondala, Chandrahas Reddy Reddy, M. Vishnu Vardhan Magha Raghul, G. Ambika, B.J. JP Morgan & Chase Co. HoustonTX United States BMS Institute of Technology and Management Dept. of AI&ML Bengaluru India Manipal Institute of Technology Bengaluru Dept. of Computer Science and Engineering Karnataka Manipal576104 India

ISBN: (纸本)9798350375466

Even with a powerful network of signaling and warning systems in the country, there have been many examples of trains crossing the red signal due to various factors, even today. These occurrences, known as Signal Passed At Danger (SPAD) events, could potentially result in severe consequences such as train derailments, train collisions, infrastructure collisions, and other dangerous events. Traditionally, these events have been analyzed using the Fault Tree Analysis (FTA) approach Bayesian Network (BN) is considered to be a better model to represent this situation when it comes to handling complexity. Bayesian Network allows for integration and includes systematic knowledge about the system, which helps to get a very flexible concise, and simple graphed representation. This study demonstrates applications of Bayesian Belief Networks (BBN) analysing and enlightening the Signal Passed at Danger (SPAD) system in railway processes. Execution of BBN offers more dynamic and flexible models compared to outmoded Fault Tree Analysis (FTA), struggling with complexity. BBN efficiently catches the inter dependencies between factors that are contributing to SPAD events, such as signal conditions, human errors, and mechanical issues. Over simulations and probability calculations using OpenMarkov software, Study revealed key intuitions, such as identifying highly risk apparatuses within SPAD system. The outcomes include enhanced risk calculation accuracy and recommendations for improving signal systems and safety protocols. These conclusions contribute to a safer and reliable railway structure, highlighting the success of BBN for complex, safety critical systems in real world applications. © 2024 IEEE.

关键词： Derailments

来源：评论

学校读者我要写书评

暂无评论

Leveraging Hallucinations to Reduce Manual Prompt Dependency in Promptable Segmentation 38

Leveraging Hallucinations to Reduce Manual Prompt Dependency...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Hu, Jian Lin, Jiayi Yan, Junchi Gong, Shaogang School of Electronic Engineering and Computer Science Queen Mary University of London United Kingdom Dept. of CSE School of AI Moe Key Lab of AI Shanghai Jiao Tong University China

Promptable segmentation typically requires instance-specific manual prompts to guide the segmentation of each desired object. To minimize such a need, task-generic promptable segmentation has been introduced, which employs a single task-generic prompt to segment various images of different objects in the same task. Current methods use Multimodal Large Language Models (MLLMs) to reason detailed instance-specific prompts from a task-generic prompt for improving segmentation accuracy. The effectiveness of this segmentation heavily depends on the precision of these derived prompts. However, MLLMs often suffer hallucinations during reasoning, resulting in inaccurate prompting. While existing methods focus on eliminating hallucinations to improve a model, we argue that MLLM hallucinations can reveal valuable contextual insights when leveraged correctly, as they represent pre-trained large-scale knowledge beyond individual images. In this paper, we utilize hallucinations to mine task-related information from images and verify its accuracy for enhancing precision of the generated prompts. Specifically, we introduce an iterative Prompt-Mask Cycle generation framework (ProMaC) with a prompt generator and a mask generator. The prompt generator uses a multi-scale chain of thought prompting, initially exploring hallucinations for extracting extended contextual knowledge on a test image. These hallucinations are then reduced to formulate precise instance-specific prompts, directing the mask generator to produce masks that are consistent with task semantics by mask semantic alignment. The generated masks iteratively induce the prompt generator to focus more on task-relevant image areas and reduce irrelevant hallucinations, resulting jointly in better prompts and masks. Experiments on 5 benchmarks demonstrate the effectiveness of ProMaC. Code given in https://***/ProMaC/. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

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：