检索结果-内蒙古大学图书馆

A Comparative Study of Deep Learning Models for Guava Leaf Disease Detection 3

A Comparative Study of Deep Learning Models for Guava Leaf D...

3rd International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies, ICAECT 2023

作者： Thangaraj, Rajasekaran Mohan, M. Moulik, M. Logeshwari, A. Loganathan, T. Koushika, P. Centre for IoT and AI Department of Computer Science and Engineering KPR Institute of Engineering and Technology Coimbatore India KPR Institute of Engineering and Technology Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9781665494007

Plant diseases are regarded as one of the major variables affecting food production and reducing production losses. Early identification and disease diagnosis of fruit crops are essential for the food industry to increase production. The development of the healthy fruit sector is highly dependent on plant diseases. Guava fruit is a good source of fiber, calcium, pectins, and vitamin C. Guava disease has become a major issue for guava production, which has a negative impact on farmers' socioeconomic growth. Infections that harm guava plants and fruits pose the greatest obstacles, although they can be avoided by early identification. Deep learning has recently been widely used to automatically identify a variety of diseases that cannot be seen with the human eye. In this study, convolution neural models are suggested to use leaf images to detect guava disease. The presented models can be used to categorize the main guava leaf diseases, including Canker, Dot, Healthy, Mummification and Rust. The kaggle dataset is used to gather images of guava with the disease Canker, Dot, Healthy, Mummification and Rust. In our work, CNN models like DenseNet121, DenseNet169, InceptionV3, and Xception are used to categorize the diseases that affect guava leaves. The experimental findings indicate that, with an accuracy of 96.12%, the DenseNet169 model outperformed other models. When the model's effectiveness is assessed using performance metrics like precision, recall, and F1 score, the results demonstrate that DenseNet169 gives excellent results. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Prediction of River Water Quality Index and Uncertainty Analysis using Machine Learning Techniques 1

Prediction of River Water Quality Index and Uncertainty Anal...

引用

1st International Conference on Software, Systems and Information Technology, SSITCON 2024

作者： Alabdeli, Haider Pareek, Piyush Kumar Shivaraj, Shruti Rupavath, Rana Veer Samara Sihman Brahmatej Das, Abhijeet The Islamic University of Al Diwaniyah Department of Computers Techniques Engineering Al Diwaniyah Iraq Nitte Meenakshi Institute of Technology Department of Ai and Ml Bengaluru India Department of Civil Engineering Raichur India Indiana Wesleyan University Department of Computer Science MarionIN United States Department of Civil Engineering Odisha Bhubaneswar India

ISBN: (纸本)9798350352931

The increasing demand for clean water in river systems drives the development of new models for water quality prediction. The problem statement is the need for accurate and reliable prediction of water quality indices in various river basins using machine learning techniques. Through the integration of diverse methodologies in the proposed approach aims to overcome limitations associated with conventional methods and emerging in water quality concerns. The dataset used in Water Quality Prediction (WQP) in the river dataset was used to predict the water quality index. The proposed method in this paper consists of preprocessing, data standardization, and data splitting with the parameter and calculating the WQI. The combined machine learning algorithms include Random Forest (RF), Extreme Gradient Boosting (XGB), and K-Nearest Neighbor (KNN). The numerical validation of this paper is an MSE of 0.03, an RMSE of 0.09, an MAE of 0.02, and R 2~ 0.12. The existing methods like Water Quality Management Limitations (WQM-WQI), Sequential Minimal Optimization-Support Vector Machine (SMO-SVM), Principal Quality Analysis (PQA-WQI), and Artificial Neural Network (ANN-WQI) are overcome by the proposed method. © 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

Analysis of Bigmart Outlets sales using Big Data Analytics Method 2

Analysis of Bigmart Outlets sales using Big Data Analytics M...

引用

2nd International Conference on Applied Artificial Intelligence and Computing, ICAaiC 2023

作者： Ramasami, Meaga Varsha Thangaraj, Rajasekaran Kumar, S Manoj Kpr Institute of Engineering and Technology Department of Computer Science and Engineering Coimbatore India Centre for IoT and Ai Kpr Institute of Engineering and Technology Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9781665456302

Sales prediction is an important problem for businesses. A sales forecast can guide a company's production, pricing, and promotion strategies, and can help mitigate risks associated with under or overproduction. Traditional methods of sales prediction utilized statistical techniques like regression analysis or time series analysis, but these methods have certain limitations, including the need for large and clean data sets. This research work focuses on exploring the patterns and trends present in a large-scale retail dataset through exploratory data analysis (EDA). The primary objective of this research is to uncover insights that can enhance the understanding of customer behaviour, identify opportunities to improve business processes and inform data-driven decisions. The data consists of daily sales records for 1560 products of ten BigMart Outlets. This study explores several preprocessing steps, such as handling missing values and outliers, scaling and normalization, and feature engineering. The findings also revealed that the top-selling items were bought by a diverse range of customer demographics, indicating that marketing campaigns should target a broader audience. Furthermore, the study identified several unusual transactions that required further investigation. © 2023 IEEE.

关键词： Sales

来源：评论

学校读者我要写书评

暂无评论

Entropy Guided Extrapolative Decoding to Improve Factuality in Large Language Models 31

Entropy Guided Extrapolative Decoding to Improve Factuality ...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Das, Souvik Jin, Lifeng Song, Linfeng Mi, Haitao Peng, Baolin Yu, Dong Department of Computer Science and Engineering University at Buffalo NY United States Tencent AI Lab BellevueWA United States

ISBN: (纸本)9798891761964

Large language models (LLMs) exhibit impressive natural language capabilities but suffer from hallucination - generating content ungrounded in the realities of training data. Recent work has focused on decoding techniques to improve factuality during inference by leveraging LLMs' hierarchical representation of factual knowledge, manipulating the predicted distributions at inference time. Current state-of-the-art approaches refine decoding by contrasting early-exit distributions from a lower layer with the final layer to exploit information related to factuality within the model forward procedure. However, such methods often assume that the final layer is the most reliable and the lower-layer selection process depends on it. In this work, we first propose the extrapolation of critical token probabilities beyond the last layer for more accurate contrasting. We additionally employ layer-wise entropy-guided lower-layer selection, decoupling the selection process from the final layer. Experiments demonstrate strong performance, surpassing state-of-the-art on multiple different datasets by large margins. The analyzes show that different kinds of prompt respond to different selection strategies. Our source code will be available on GitHub. © 2025 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

BiGRU-ANN based hybrid architecture for intensified classification tasks with explainable ai

引用

International Journal of Information Technology (Singapore) 2023年第8期15卷 4211-4221页

作者： Chakraborty, Sovon Talukder, Muhammad Borhan Uddin Hasan, Mohammad Mehadi Noor, Jannatun Uddin, Jia Department of Computer Science and Engineering University of Liberal Arts Bangladesh Dhaka Bangladesh Department of Computer Science and Engineering Daffodil International University Dhaka Bangladesh Department of Computer Science and Engineering European University of Bangladesh Dhaka Bangladesh Department of Computer Science and Engineering BRAC University Dhaka Bangladesh AI and Big Data Department Woosong University Daejeon South Korea

Artificial Intelligence (ai) is increasingly being employed in critical decision-making processes such as medical diagnosis, credit approval, criminal justice, and many more. However, many ai models exploit complex algorithms that are difficult for humans to see through, which can lead to concerns about accountability, bias, and the ability to trust the outcomes. With the increasing demand for ai systems to be transparent, interpretable, and reliable, the field of Explainable ai (Xai) has gained attention of the researchers. This paper presents a robust hybrid architecture that combines Bidirectional Gated Recurrent Units (BiGRU) and Artificial Neural Networks (ANN) for the classification of texts and sentiment analysis. Interpretable Model Agnostic Explanation (LIME) has been employed with our proposed model to enhance confidence in the outcomes. The proposed architecture is found to be effective for sentiment analysis from texts, and classifying images containing handwrit- ten characters. It leverages the BiGRU to model the sequential dependencies in the data, while the ANN is used for the final classification. Evaluations on both Bengali and English datasets show that the proposed architecture outperforms state-of-the-art models in various performance metrics, providing meaningful and interpretable explanations for its predictions. The model can be used in systems that require the architectures to be computationally less demanding, yet a decent accuracy is secured. © 2023, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of computer Applications and Management.

关键词： ANN BiGRU Explainable ai LIME Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

The Effect of Programming Language in Software Bug Prediction 3

The Effect of Programming Language in Software Bug Predictio...

引用

3rd International Conference on Communication, Computing and Industry 4.0, C2I4 2022

作者： Jayadev, Gopika Gayathri, Kaparthi Babu, Tina Gandiboina, Harika Sudha, Kavya Venkteswalru, Bondu School of Engineering Dayananda Sagar University Department of AI and ML Bengaluru India School of Engineering Dayananda Sagar University Department of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9781665474979

Software maintenance is a very important phase in the life cycle of software development. As part of maintenance, we need to identify bugs within the code and fix them for every release. Software bug prediction (SBP) allows us to identify modules in the software that may have the tendency to be buggy. This enables us to perform targeted testing and properly plan maintenance cycles. In most research performed, we observed that dataset of software programs written in C language were used. Each programming language are inherently different and has it's own constructs, practices and nuances. In this research we focus on identifying if there exists a specific classifier that works well for each programming language which in turn is used to analyze the effect of programming language on Software Bug Prediction. Datasets of software written in C, C++ and Java has been collected and the most accurate classifier for each dataset of a programming language has been identified. The ml models used in this paper include Naive Bayes, Decision Tree and Random Forest. © 2022 IEEE.

关键词： Decision Tree Machine Learning Naïve Bayes Random Forest Software Bug Prediction

来源：评论

学校读者我要写书评

暂无评论

Next-Gen Human-computer Interaction: A Hybrid LSTM-CNN Model for Superior Adaptive User Experience

Next-Gen Human-Computer Interaction: A Hybrid LSTM-CNN Model...

引用

Electrical, Electronics, Information and Communication Technologies (ICEEICT), International Conference on

作者： Ravindra Changala J. Manoranjini M. Misba Shital Kakad S Sai Deepthi Ponni Valavan M Department of Computer Science and Engineering Guru Nanak Institutions Hyderabad India Department Of AI&ML Rajalakshmi Engineering College Chennai India Department CSE (AI&ML) Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Chennai Tamil Nadu India Department of AI & DS Vishwakarma Institute of Technology Pune Maharashtra Department IT B V Raju Institute of Technology (Autonomous) Narsapur Hyderabad India Department of AI & DS K. Ramakrishnan College of Engineering Trichy Tamil Nadu India

ISBN: (数字)9798350369083

ISBN: (纸本)9798350369090

In the rapidly evolving landscape of human-computer interaction (HCI), the demand for personalized and adaptive user experiences has grown exponentially. To meet this demand, Research propose a groundbreaking approach leveraging the fusion of Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) architectures. This hybrid LSTM-CNN model is designed to enhance adaptability, responsiveness, and user engagement across various interactive platforms. Traditional HCI models often struggle to effectively capture the dynamic nature of user behavior and preferences. By integrating LSTM and CNN, This model achieves a synergistic blend of temporal and spatial feature extraction capabilities. The LSTM component excels in capturing sequential dependencies and long-term patterns, enabling the system to learn from past interactions and anticipate future user actions. Meanwhile, the CNN component efficiently processes spatial information, extracting meaningful features from multimedia inputs such as images, videos, and text. One of the key strengths of proposed model is its adaptability to diverse user contexts and preferences. Through continuous learning and adaptation, it dynamically adjusts its interface, content, and interaction patterns to match the evolving needs and preferences of individual users. Moreover, the hybrid architecture enables real-time processing of multimodal inputs, facilitating seamless interaction across a wide range of devices and platforms. In experimental evaluations, Hybrid LSTM-CNN model demonstrated superior performance compared to baseline methods in terms of user satisfaction, engagement, and task completion rates. Furthermore, it exhibited robustness and scalability, making it suitable for deployment in real-world applications across domains such as e-commerce, entertainment, education, and healthcare. In summary, The proposed approach represents a significant advancement in HCI research, paving the way for next-generation interactiv

关键词： Human computer interaction Adaptation models Scalability Streaming media Feature extraction User experience Robustness Convolutional neural networks Long short term memory Videos

来源：评论

学校读者我要写书评

暂无评论

Leveraging Machine Learning for Pattern-Based Detection of Respiratory Illnesses

Leveraging Machine Learning for Pattern-Based Detection of R...

引用

Artificial Intelligence and Data engineering (aiDE), International Conference on

作者： Anusha Prashanth Shetty Deepthi L Raksha Puthran Surendra Shetty Pavan Hegde Shravan Kumar Department of Master of Computer Applications NMAM Insitute of Technology NITTE (Deemed to be University) Nitte Karkala India Department of Computer Science and Engineering (AI&ML) Sahyadri College of Engineering and Management Mangalore India Department of Paediatrics Father Muller Medical College and Hospital Mangalore India

ISBN: (数字)9798331527518

ISBN: (纸本)9798331527525

Lung sound analysis presents a promising, noninvasive, and potentially cost-effective approach for detecting respiratory diseases. This research aims to assess the feasibility of using lung sounds for pattern recognition and detection of various respiratory disorders, specifically by conducting an analysis of cough sounds. We review a range of recent studies that explore the use of deep learning models, machine learning algorithms, and feature extraction techniques in this context. These studies are categorized based on their methodologies, and their performance in accurately classifying lung sounds is discussed. Research findings highlight the significant diagnostic potential of lung sound analysis, demonstrating improved classification accuracies across various disease types. However, challenges such as data inconsistency, noise interference, and limited model generalizability remain. It captures the current state of lung sound analysis research in respiratory disease detection, offering insights that could drive future advancements in this critical field.

关键词： Machine learning algorithms Accuracy Tuberculosis Lungs Pulmonary diseases Medical services User interfaces Data models Convolutional neural networks Mel frequency cepstral coefficient

来源：评论

学校读者我要写书评

暂无评论

Faithfulness Measurable Masked Language Models 41

Faithfulness Measurable Masked Language Models

引用

41st International Conference on Machine Learning, ICml 2024

作者： Madsen, Andreas Reddy, Siva Chandar, Sarath Mila Montreal Canada Computer Engineering and Software Engineering Department Polytechnique Montreal Montreal Canada Computer Science and Linguistics McGill University Montreal Canada Facebook CIFAR AI Canada Canada CIFAR AI Canada

A common approach to explaining NLP models is to use importance measures that express which tokens are important for a prediction. Unfortunately, such explanations are often wrong despite being persuasive. Therefore, it is essential to measure their faithfulness. One such metric is if tokens are truly important, then masking them should result in worse model performance. However, token masking introduces out-of-distribution issues, and existing solutions that address this are computationally expensive and employ proxy models. Furthermore, other metrics are very limited in scope. This work proposes an inherently faithfulness measurable model that addresses these challenges. This is achieved using a novel fine-tuning method that incorporates masking, such that masking tokens become in-distribution by design. This differs from existing approaches, which are completely model-agnostic but are inapplicable in practice. We demonstrate the generality of our approach by applying it to 16 different datasets and validate it using statistical in-distribution tests. The faithfulness is then measured with 9 different importance measures. Because masking is in-distribution, importance measures that themselves use masking become consistently more faithful. Additionally, because the model makes faithfulness cheap to measure, we can optimize explanations towards maximal faithfulness;thus, our model becomes indirectly inherently explainable. Copyright 2024 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

PAC PREDICTION SETS UNDER LABEL SHIFT 12

PAC PREDICTION SETS UNDER LABEL SHIFT

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Si, Wenwen Park, Sangdon Lee, Insup Dobriban, Edgar Bastani, Osbert Department of Computer & Information Science University of Pennsylvania United States Graduate School of AI and Computer Science & Engineering POSTECH Korea Republic of Department of Statistics and Data Science University of Pennsylvania United States

Prediction sets capture uncertainty by predicting sets of labels rather than individual labels, enabling downstream decisions to conservatively account for all plausible outcomes. Conformal inference algorithms construct prediction sets guaranteed to contain the true label with high probability. These guarantees fail to hold in the face of distribution shift, which is precisely when reliable uncertainty quantification can be most useful. We propose a novel algorithm for constructing prediction sets with PAC guarantees in the label shift setting, where the probabilities of labels can differ between the source and target distributions. Our algorithm relies on constructing confidence intervals for importance weights by propagating uncertainty through a Gaussian elimination algorithm. We evaluate our approach on four datasets: the CIFAR-10 and ChestX-Ray image datasets, the tabular CDC Heart Dataset, and the AGNews text dataset. Our algorithm satisfies the PAC guarantee while producing smaller prediction set sizes compared to several baselines. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Forecasting

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：