检索结果-内蒙古大学图书馆

Robust Multi-Dialect End-to-End ASR Model Jointly with Beam Search Threshold Pruning and LLM

SN Computer science 2025年第4期6卷 1-12页

作者： Shunmuga Priya, M.C. Karthika Renuka, D. Ashok Kumar, L. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India Department of Information Technology PSG College of Technology Coimbatore India Department of Electrical and Electronics Engineering Thiagarajar College of Engineering Madurai India

This paper aims to develop a novel robust multi-dialect end-to-end ASR system with beam search threshold pruning. The efficacy of our proposed model is evaluated using word error rate (WER). Our key contributions are: (1) To develop an end-to-end ASR system using attention-based neural network architecture and analyze the effectiveness of two features such as MFCC and log mel filter bank energies on multiple speech dialect corpora including American, Britain, and Indian accents;(2) To integrate beam search threshold pruning as a decoding mechanism to reduce the decoding time (3) To conduct an experimental analysis to test the model performance and compare the results against baseline system. (4) Post processing analysis are carried out using Llama2-7B based large language model(LLM) for enhancing the performance of proposed ASR system. The proposed model significantly improves performance by 1.91% and 4.29% over clean and noisy speech in librispeech corpus. Similarly, for the Indian accented speech, the model attains an average WER of about 6.6%. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Automatic speech recognition Beam search Decoding LLM Log Mel filter bank energies

来源：评论

学校读者我要写书评

暂无评论

Deep Convolutional Neural Networks for Early-Stage Detection and Prognostication of Lung and Colon Cancer 2

Deep Convolutional Neural Networks for Early-Stage Detection...

引用

2nd IEEE International Conference on Integrated Circuits and Communication Systems, ICICACS 2024

作者： Laxmikant, Kumar Arthi, A. Vinodhini, V. Natarajan, B. Bhuvaneswari, R. Selvam, Prabu Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Chennai India Karpagam College of Engineering Computer Science and Technology Coimbatore India University College of Engineering Department of Electronics and Communication Engineering Ariyalur India VIT University School of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350317558

Lung and colon cancers are significant health issues across the world, prompting the need for inventive methods of diagnosis. This study takes the lead in introducing advanced Deep ConvNets (CNNs) to enhance the accuracy of early detection. The model is trained on extensive datasets, resulting in impressive outcomes. During training, it achieves an accuracy of 92.54% and a loss value of 0.0161. These results are mirrored in testing, where the accuracy reaches 96.59% and the loss value is 0.0441. Beyond just numerical achievements, this research revolutionizes cancer diagnostics by providing a sophisticated tool for personalized and early detection. This study surpasses the mere influence on individuals and extends its reach to bring hope for society and the economy by tackling the pervasive weight of cancer. It distinguishes itself in the realm of advanced medical investigation, affirming the efficacy of the model and inspiring additional inquiry, encompassing the utilization of more extensive sets of data and implementation in diverse medical domains. © 2024 IEEE.

关键词： Colon Cancer Convolutional Neural Network Lung Cancer

来源：评论

学校读者我要写书评

暂无评论

Cascaded PFLANN Model for Intelligent Health Informatics in Detection of Respiratory Diseases from Speech Using Bio-inspired Computation

引用

Journal of Artificial Intelligence and Technology 2024年第2期4卷 124-131页

作者： Jagannath Dayal Pradhan L.V.Narasimha Prasad Tusar Kanti Dash Manisha Guduri Ganapati Panda Electronics and Communications Engineering C V Raman Global UniversityBhubaneswarIndia Department of Computer Science and Engineering Institute of Aeronautical EngineeringHyderabadIndia School of Computing and Informatics University of LouisianaLafayetteLAUSA

Due to the recent developments in communications technology,cognitive computations have been used in smart healthcare techniques that can combine massive medical data,artificial intelligence,federated learning,bio-inspired computation,and the Internet of Medical *** has helped in knowledge sharing and scaling ability between patients,doctors,and clinics for effective treatment of ***-based respiratory disease detection and monitoring are crucial in this direction and have shown several promising *** the subject’s speech can be remotely recorded and submitted for further examination,it offers a quick,economical,dependable,and noninvasive prospective alternative detection ***,the two main requirements of this are higher accuracy and lower computational complexity and,in many cases,these two requirements do not correlate with each *** problem has been taken up in this paper to develop a low computational complexity-based neural network with higher accuracy.A cascaded perceptual functional link artificial neural network(PFLANN)is used to capture the nonlinearity in the data for better classification performance with low computational *** proposed model is being tested for multiple respiratory diseases,and the analysis of various performance matrices demonstrates the superior performance of the proposed model both in terms of accuracy and complexity.

关键词： AI-ML ASR FLANN health informatics neural network PFLANN

来源：评论

学校读者我要写书评

暂无评论

Enhancing Agricultural IoT with Predictive Analytics and Reinforcement Learning Models 2

Enhancing Agricultural IoT with Predictive Analytics and Rei...

引用

2nd IEEE International Conference on Integrated Circuits and Communication Systems, ICICACS 2024

作者： Krishnan, R. Santhana Raj, J. Relin Francis Christopher, V. Bibin Nandini, N. Muthu, A. Essaki Puvanadevi, C. SCAD College of Engineering and Technology Department of Electronics and Communication Engineering Cheranmahadevi India Ssveetha School of Engineering Saveetha Institute of Medical and Technical Science Department of Electronics and Communication Engineering Chennai India Faculty of Engineering and Technology SRM Institute of Science and Technology Department of Computing Technologies School of Computing College of Engineering and Technology Kattankulathur Chennai India RajaRajeswari College of Engineering Department of Electrical and Electronics Engineering Bangalore India SCAD College of Engineering and Technology Department of Computer Science and Engineering Cheranmahadevi India Adhiyamaan College of Engineering Department of Information Technology Hosur India

ISBN: (纸本)9798350317558

This research explores the augmentation of Agricultural Internet of Things (IoT) systems through the integration of advanced predictive analytics and reinforcement learning models. A novel algorithm, termed "CropQL,"is proposed to optimize crop yield prediction and resource allocation in precision agriculture. CropQL amalgamates the strengths of the CropNet IoT model with Q-Learning for reinforcement learning and Random Forest for predictive analytics. Simulation analyses are conducted to evaluate the performance of CropQL against established algorithms, employing diverse metrics such as precision, recall, F1 score, and area under the receiver operating characteristic curve (AUC-ROC). The proposed CropQL algorithm exhibits superior predictive accuracy, demonstrating a significant enhancement in crop yield forecasting when compared to baseline algorithms. Precision and recall metrics reveal the algorithm's efficacy in minimizing false positives and negatives, ensuring precision agriculture systems make informed decisions. F1 score indicates the balanced optimization of precision and recall, reinforcing the algorithm's robustness. AUC-ROC analysis further corroborates the superior discriminative power of CropQL in distinguishing between positive and negative instances. In comparison to existing algorithms, including AgriSensNet, PrecisionCropNet, and IoTFarmGuard, CropQL consistently outperforms in predictive accuracy and resource optimization. The simulation results underscore the efficacy of the proposed algorithm in enhancing the efficiency of agricultural IoT systems, providing a promising avenue for sustainable precision farming. © 2024 IEEE.

关键词： Agricultural IoT CropQL Precision Agriculture Predictive Analytics Reinforcement Learning Sustainable Farming

来源：评论

学校读者我要写书评

暂无评论

A Deep CNN-Based Approach for 10-Class with Two-Channel EMG Signal Classification 4th

A Deep CNN-Based Approach for 10-Class with Two-Channel EMG ...

引用

4th International Conference on electronics, Biomedical engineering, and Health Informatics, ICEBEHI 2023

作者： Triwiyanto Yulianto, Endro Rahmawati, Triana Chai, Rifai Department of Medical Electronics Technology Poltekkes Kemenkes Surabaya Surabaya Indonesia School of Science Computing and Engineering Technologies Swinburne University of Technology Melbourne Australia

ISBN: (纸本)9789819714629

Electromyography (EMG) signals have emerged as vital tools for prosthetic control, motor function assessment, and rehabilitation technology. Traditional methods, relying on manual feature extraction, struggle with high-dimensional EMG data. Deep learning, specifically Convolutional Neural Networks (CNNs), has shown promise in addressing these challenges. However, there is a need to explore CNNs’ effectiveness, optimal architecture, real-time processing, and data augmentation for EMG signal classification. This study aimed to introduce a CNN-based approach tailored to classify ten-class, two-channel EMG signals. Eight participants provided EMG data by performing ten distinct finger and hand movements. Data were collected using two-channel EMG sensors, amplified, sampled at 4000 Hz, and filtered. A CNN architecture with eight layers was designed, incorporating one-dimensional convolutional layers, max pooling, global average pooling, dropout layers, and a dense output layer. The CNN-based approach demonstrated promising results. It exhibited high accuracy in classifying EMG signals for most classes, with an overall accuracy of approximately 91%. It excelled in the thumb (THU) and thumb-ring (TH-R) classes. However, some classes, like thumb-middle (TH-M) and thumb-index (TH-I), showed room for improvement in terms of recall. This study introduced an innovative CNN-based approach to tackle the challenges of classifying ten-class, two-channel EMG signals. It addressed critical research gaps and significantly advanced EMG signal processing, particularly for prosthetic control and rehabilitation technology. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

A Novel Predictive Analysis and Classification of Land Subsidence Vulnerability Mapping based on GIS using Hybrid Optimized Machine Learning Techniques and Computer Vision 5

A Novel Predictive Analysis and Classification of Land Subsi...

引用

5th International Conference on Innovative Data Communication Technologies and Application, ICIDCA 2024

作者： Senthil, G.A. Prabha, R. Arun, M. Sahaya Anselin Nisha, A. Prabu, S. Department of Information Technology Agni College of Technology Chennai India Department of Electronics and Communication Engineering Sri Sairam Institute of Technology Chennai India Department of Electronics and Communication Engineering Panimalar Engineering College Chennai India Department of Electronics and Communication Engineering Sathyabama Institute of Science and Technology Chennai India Department of Computing Technologies School of Computing SRM Institute of Science and Technology Kattankulathur Chennai India

Land subsidence is a natural disaster caused by the extraction of the earth's material strength on the surface that causes the earth's surface to settle and sink. Land subsidence can also result in significant financial losses since it causes structural damage and expensive maintenance costs for roads, trains, barriers, pipelines, ground water, oil, natural gas, mines, and buildings. The Geographic Information System (GIS) and Google Earth Platform (GEP) can be used to map this land subsidence. This provides a comprehensive view of land pattern subsidence and its interactions with other geographical elements. It applies to both progressive subsidence caused by fluid extraction (groundwater, oil, or natural gas) and rapid surface subsidence caused by the collapse of an underground mine. Risk mitigation is the most effective way to deal with the experimental model. This research study focuses on the use of Predictive Data Analysis (PDA), a powerful environment in data analytics that provides a set of tools that make it easier to evaluate optimal models. The proposed model makes use of a dataset that includes specific land subsidence circumstances as input. The prediction result is achieved using a hybrid machine learning technique that combines Stochastic Gradient Descent (SGD) with Long Short-Term Memory (LSTM) and is compared to K-Nearest Neighbor (KNN), Generative Adversarial Network (GAN), and Artificial Neural Network (ANN). The proposed system employs LSTM and SGD as optimizers, resulting in a hybrid model that improves accuracy. Among the four algorithms investigated the LSTM-SGD network has the highest accuracy (97%), followed by KNN (96%), GAN (90%), and ANN (89%). Due to their ability to capture complicated nonlinear correlations in data and manage temporal dependencies, optimized hybrid ML models LSTM-SGD and KNN models appear to be particularly well-suited for land subsidence prediction. © 2024 The Authors. Published by Elsevier B.V.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

A Grading Model for Jasmine Flowers Using Deep Learning and Quantum Machine Learning 4

A Grading Model for Jasmine Flowers Using Deep Learning and ...

引用

4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Pratiksha Cauvery, K.P. Thrisha, R. Harshitha, T. Panigrahi, Suman Thangavel, Senthil Kumar Somasundaram, K. Boopathy, C.P. Amrita School of Physical Sciences Department of Mathematics Tamil Nadu Coimbatore India Amrita School of Computing Department of Computer Science and Engineering Tamil Nadu Coimbatore India Amrita School of Engineering Department of Elelctical and Electronics Engineering Tamil Nadu Coimbatore India

ISBN: (纸本)9798331523923

The flowers of Jasmine hold immense economic importance in the agricultural sector and the perfume industry, as the value and satisfaction of the market, are heavily influenced by the quality of these flowers. The already very labor-intensive, subjective, and inefficient manual method of grading plants based on their physical and aromatic characteristics is unsuitable on a large scale. This paper handles these limitations by introducing automation in the grading process using Deep Learning and Quantum Machine Learning. The conceptual methodology mandates automatic flower placement and high-resolution imagery with a conveyor belt. The deep learning model, exactly a type of Convolutional Neural Network (CNN), would explain freshness in terms of visual features like color, texture, structural strength, and vice versa. Additionally, quantum-inspired algorithms have been employed to enhance the computation and performance of the model. From the analysis, a decision-making module classifies entrant flowers into fresh and otherwise, thereby sending them to their respective places, where all packing or disposal could be achieved. This DL-QML hybrid approach points out the exemplary high accuracy, scalability, and efficiency in jasmine grading and hence minimizes manual intervention in quality control assembly lines. The findings provide the card of the feasibility of joining QML in addressing computational challenges of agro-product quality assessment, opening avenues in application far broader than floriculture. © 2025 IEEE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Adaptive Genghis Khan Shark based convolutional denoising autoencoder for quality enhancement of satellite images

引用

International Journal of Information Technology (Singapore) 2025年 1-14页

作者： Rajalakshmi, M. Subramanian, R. S. Sankara Tamilselvan, S. Bhuvanesh, A. Department of Computing Technologies School of Computing SRM Institute of Science and Technology Tamil Nadu Chennai India Department of Mathematics PSG Institute of Technology and Applied Research Tamil Nadu Coimbatore India Department of Electronics and Communication Engineering Sri Eshwar College of Engineering Tamil Nadu Coimbatore India Department of Electrical and Electronics Engineering PSN College of Engineering and Technology Tamil Nadu Tirunelveli India

Satellite image processing often enhances feature perception and visibility for exact feature identification. Additionally, the change improves remote sensing data quality and clarity. Due to their long distances, these photos often have weather-related noise and distortion that must be fixed. This paper addresses these challenges with a Deep Learning (DL) satellite image denoising technique. An Adaptive Genghis Khan Shark Algorithm (AGKSA) and a Convolutional Denoised Autoencoder are used to improve satellite images in this study. The CDAE settings are optimized using the AGKSA while keeping important image attributes. AGKSA is used in this hybrid model to optimize CDAE performance under various noise conditions. Experimental results showed that the proposed model outperforms traditional image denoising approaches in satellite imagery feature extraction and visual quality. In very precise image processing applications like environmental monitoring and land use classification, the CDAE-AGKSA works well. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2025.

关键词： Adaptive Genghis Khan algorithm Convolutional denoised autoencoder Satellite image enhancement Satellite image processing

来源：评论

学校读者我要写书评

暂无评论

Dementia Detection Using Machine Learning in Males and Females

Dementia Detection Using Machine Learning in Males and Femal...

引用

2024 International Conference on Artificial Intelligence and Emerging Technology, Global AI Summit 2024

作者： Iyer, Aishwarya Hariharan Vekkot, Susmitha Gupta, Deepa Gopalakrishnan, E.A. School of Computing National University of Singapore Singapore Amrita School of Engineering Amrita Vishwa Vidyapeetham Dept. of Electronics and Communication Engineering Bangalore India Amrita School of Computing Amrita Vishwa Vidyapeetham Dept. of Computer Science and Engineering Bangalore India Amrita School of Artificial Intelligence Amrita Vishwa Vidyapeetham Bangalore India

ISBN: (纸本)9798350379716

In this study, we train different Machine Learning classifiers on male and female samples from The ADReSS Challenge dataset. After testing them on data consisting of both male and female samples as well as on data with samples of only the corresponding sex, we use SHapley Additive exPlanations on the models to identify the most important features. Then, we present the results of training the models with only the most important features identified. Finally, we present a comparative analysis based on the observations and insights gathered through these experiments. © 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

Three-Stage Transfer Learning with AlexNet50 for MRI Image Multi-Class Classification with Optimal Learning Rate

引用

Computer Modeling in engineering & sciences 2025年第1期142卷 155-183页

作者： Suganya Athisayamani A.Robert Singh Gyanendra Prasad Joshi Woong Cho School of Computing Sastra Deemed to be UniversityThanjavur613401India Department of Computational Intelligence SRM Institute of Science and TechnologyKattankulathur603203India Department of AI and Software Engineering Kangwon National UniversitySamcheok25913Republic of Korea Department of Electronics Information and Communication EngineeringKangwon National UniversitySamcheok25913Republic of Korea

In radiology,magnetic resonance imaging(MRI)is an essential diagnostic tool that provides detailed images of a patient’s anatomical and physiological *** is particularly effective for detecting soft tissue ***,radiologists manually interpret these images,which can be labor-intensive and time-consuming due to the vast amount of *** address this challenge,machine learning,and deep learning approaches can be utilized to improve the accuracy and efficiency of anomaly detection in MRI *** manuscript presents the use of the Deep AlexNet50 model for MRI classification with discriminative learning *** are three stages for learning;in the first stage,the whole dataset is used to learn the *** the second stage,some layers of AlexNet50 are frozen with an augmented dataset,and in the third stage,AlexNet50 with an augmented dataset with the augmented *** method used three publicly available MRI classification datasets:Harvard whole brain atlas(HWBA-dataset),the school of Biomedical engineering of Southern Medical University(SMU-dataset),and The National Institute of Neuroscience and Hospitals brain MRI dataset(NINS-dataset)for *** hyperparameter optimizers like Adam,stochastic gradient descent(SGD),Root mean square propagation(RMS prop),Adamax,and AdamW have been used to compare the performance of the learning ***-dataset registers maximum classification *** evaluated the performance of the proposed classification model using several quantitative metrics,achieving an average accuracy of 98%.

关键词： MRI tumors classification AlexNet50 transfer learning hyperparameter tuning optimizer

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：