检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Wang, Wei Qian, Yanmin Shanghai Jiao Tong University MoE Key Lab of Artificial Intelligence Ai Institute X-LANCE Lab Department of Computer Science and Engineering Shanghai China

ISBN: (纸本)9781728163277

Self-supervised learning (SSL) has attracted widespread research interest since many successful SSL approaches such as wav2vec 2.0 and Hidden-unit BERT (HuBERT) have achieved promising results on speech-related tasks such as automatic speech recognition (ASR). However, few works have been conducted to improve the noise robustness of SSL models. In this paper, we propose HuBERT-AGG, a novel method that learns noise-invariant SSL representations for robust speech recognition by distilling aggregated layer-wise representations. Specifically, we learn an aggregator that computes the weighted sum of all hidden states of a pretrained vanilla Hu-BERT by fine-tuning it on a small portion of labeled data. Then a noise-robust HuBERT is trained on the simulated noisy speech by distilling from the aggregated representations and layer-wise hidden states produced by a pretrained vanilla HuBERT with parallel original speech as input. Experiments on libriSpeech simulated noisy test sets show 13.1%-17.0% relative word error rate (WER) reduction with very slight degradation on the original test sets. On CHiME-4 1-channel real speech test sets, we have surpassed the best results achieved by all published fully supervised ASR models as well as other SSL approaches adopting the same data usage as ours. © 2023 IEEE.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

The Influence of Mobile Commerce on Consumer Behavior: A FCM-RF-DNN Analysis

The Influence of Mobile Commerce on Consumer Behavior: A FCM...

引用

2024 International Conference on Intelligent Algorithms for Computational Intelligence Systems, IACIS 2024

作者： Akram, Nawab Dahiya, Suman Chittapragada, Hemanand Periyasamy, V.M. Vidhya, P. Chauhan, Amit Magadh Professional Institute Patna India Universal Ai University Department of Management Karjat India Madanapalle Institute of Technology & Science Computer Science & Engineering Madanapalle India Sona College of Technology Department of Eee Salem India M.Kumarasamy College of Engineering Department of Artificial Intelligence and Data Science Karur India Department of Life Sciences Bengaluru India

ISBN: (纸本)9798350360660

For m-commerce vendors, the difficulty is to decipher what factors impact customer actions in the ubiquitous mobile setting. In addition, companies are attempting to incorporate social media into their mobile approach in some way. This proposed approach to the findings of a qualitative exploratory study regarding the use of social media and smartphones within the framework of mobile commerce. Keep in mind the order of importance while doing data preprocessing, feature selection, and training the model. The usual steps in getting data ready for processing, such as cleaning it, identifying users and sessions, and finding episodes. The IS-DT suggested method's implementation technique is utilized in feature selection. Unified FCM-RF-DNN models need to be trained after features have been retrieved. Two state-of-the-art approaches, RF and DNN, are outperformed by the suggested approach. Following the implementation of the method, accuracy improved by 96.13%. © 2024 IEEE.

关键词： Mobile commerce

来源：评论

学校读者我要写书评

暂无评论

Mitigating Catastrophic Forgetting in Deep Learning Models for Sentiment Analysis

Mitigating Catastrophic Forgetting in Deep Learning Models f...

引用

International Conference on Advances in Information Technology (ICaiT)

作者： J. Ranjith Santhi Baskaran B. Adithya Department of Computer Science and Engineering Puducherry Technological University Puducherry India Department of Inofrmation Technology Puducherry Technological University Puducherry India Department of CSE(AI&ML) Geethanjali College of Engineering and Technology Hyderabad India

ISBN: (数字)9798350383867

ISBN: (纸本)9798350383874

Deep learning models have been employed to evaluate sentiment analysis. However, the issue of catastrophic forgetting is frequently seen in these models. Cognitive decline occurs when the model's performance on previously learned tasks deteriorates as it acquires new ones. A study is conducted to investigate the impact of catastrophic forgetting on sentiment analysis algorithms and propose potential solutions to address this issue. This study evaluates the adaptability of Elastic Weight Consolidation (EWC) and Learning without Forgetting (LwF) in effectively handling novel sentiment analysis tasks while preserving the model's performance on previously acquired tasks. Experimental results on the Amazon Product Reviews dataset demonstrate that both EWC and LwF significantly reduce the impact of catastrophic forgetting, with LwF outperforming EWC in terms of sentiment classification accuracy. Our findings highlight the importance of addressing catastrophic forgetting in sentiment analysis models to ensure their robustness and reliability in real-world applications.

关键词： Deep learning Training Sentiment analysis Analytical models Adaptation models Accuracy Reviews Robustness Data models Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Prediction of Weather Forecasting with Long Short-Term Memory using Deep Learning 4

Prediction of Weather Forecasting with Long Short-Term Memor...

引用

4th International Conference on Smart Electronics and Communication, ICOSEC 2023

作者： Karthika, S. Priyanka, T. Indirapriyadharshini, J. Sadesh, S. Rajeshkumar, G. Rajesh Kanna, P. Intelligence and Machine Learning Bannari Amman Institue of Technology Department of Artificial Sathyamangalam India M.Kumarasamy College of Engineering Department of Computer Science and Engineering Karur India Sri Krishna College of Engineering and Technology Department of Mechatronics Engineering Coimbatore India Velalar College of Engineering and Technology Department of Ai and Ds Erode India Bannari Amman Institute of Technology Department of Computer Science and Engineering Sathyamangalam India

ISBN: (纸本)9798350300888

Weather forecasting has a greater effect on human life, recently this filed has gained a significant research attention. The existing numerical weather prediction models are very complex to solve. Therefore, the primary goal of this research work is to create a novel and lightweight weather forecasting model. Here, the time series data have been used from Kaggle website. The obtained dataset is then processed with the help of deep learning techniques. The Long Short-Term Memory (LSTM) algorithm is used to produce better result with more accuracy when compared with other deep learning methodologies. Therefore, from the results it is evident that the proposed model can be used to predict the weather conditions with greater accuracy and assist farmers and others who depend on weather. © 2023 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Enhancing Patient Care: A Data-Driven Approach Using Decision Tree-ELM Models for Improved Safety and Quality

Enhancing Patient Care: A Data-Driven Approach Using Decisio...

引用

Advancement in Technology (ICONAT), International Conference for

作者： Supriya Shrikant Telsang Akhya Sanjay Singh Simarpreet Kaur Bramah Hazela Shikha Singh Hemanand Chittapragada Vishwakarma Institute of Technology Bibwewadi Pune Maharashtra India Computer Engineering (AI & ML) Shah & Anchor Kutchhi Engineering College Mumbai India Department of Computer Science and Engineering Amity School of Engineering and Technology Lucknow Amity University Uttar Pradesh India Department of Computer Science & Engineering Madanapalle Institute of Technology & Science Madanapalle Andhra Pradesh India

ISBN: (数字)9798350354171

ISBN: (纸本)9798350354188

In recent years, educational initiatives that seek to combine teaching medical students about patient safety and quality improvement (QI) have received funding from both the World Health Organization and the American Association of Medical Colleges. When developing curriculum, it is important to keep in mind the interests and values of the target audience. The proposed used a poll to find out how students felt about lessons on patient safety and quality improvement. It would be more efficient to incorporate patient safety and quality improvement into clinical education rather than teach it in separate computer modules or during pre-clinical lectures. This proposed approach consists of three stages, which are feature selection, model training, and preprocessing. Before doing thorough text processing, stop words are typically filtered out in conventional methods of natural language processing and information retrieval. Unigram and bigram selection are part of feature selection. During the entire model training process, a Decision Tree-ELM Ensemble was utilized. This innovative approach outperforms ELM and Decision Tree with an average accuracy of ${9 4 . 3 6 \%}$.

关键词： Training Accuracy Medical services Feature extraction Information filters Natural language processing Safety Decision trees Systems thinking Text processing

来源：评论

学校读者我要写书评

暂无评论

A Powerful Algorithm for e-Commerce Credit Risk Analysis

A Powerful Algorithm for e-Commerce Credit Risk Analysis

引用

2023 International Conference on Data science and Network Security, ICDSNS 2023

作者： Dayanand Lal, N. Guruswamy, Sandhya Lalitha, Vuyyuru Awasthy, Mohan Pareek, Piyush Kumar Gitam University Bengaluru Campus Dept. of Cse Gst Bengaluru India School of Computing and Information Technology Reva University India Koneru Lakshmaiah Education Foundation Department of Computer Science and Engineering Vaadeswaram India Bharati Vidyapeeth Peeth Deemed to Be University Department of Engineering and Technology Navi Mumbai India Nitte Meenakshi Institute of Technology Department of Ai and Ml Bengaluru India

ISBN: (纸本)9798350301595

It is critical to build a system and model for analysing the credit risk associated with e-commerce to decrease the chance of credit defaults occurring during online transactions. This approach examines a client's income, assets, negative transaction rate, reputation, and third-party credit evaluation findings to determine creditworthiness. The risk assessment index-based credit risk analysis model is built using the RB-XGBoost approach. The credit risk analysis algorithm's training process employs a data balancing technique to increase the system's capacity to forecast accurately. This method is designed to solve the problem of class imbalance in machine learning datasets. The training set is separated into groups at this point, and each group is compared to a specified threshold. The dataset is then made more representative by combining oversampling and undersampling. The system's capacity to analyse the credit risk presented by online shops is improved by balancing data subsets. Online retailers might utilise the proposed model and indexing approach to better assess their clients' creditworthiness and thereby reduce the number of consumers whose payments are late. By suggesting a unique way to address class imbalance in machine learning datasets, this study adds to the expanding body of work on credit risk assessment in online transactions. © 2023 IEEE.

关键词： Risk assessment

来源：评论

学校读者我要写书评

暂无评论

CoE-SQL: In-Context Learning for Multi-Turn Text-to-SQL with Chain-of-Editions

CoE-SQL: In-Context Learning for Multi-Turn Text-to-SQL with...

引用

2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2024

作者： Zhang, Hanchong Cao, Ruisheng Xu, Hongshen Chen, Lu Yu, Kai X-LANCE Lab Department of Computer Science and Engineering MoE Key Lab of Artificial Intelligence SJTU AI Institute Shanghai Jiao Tong University Shanghai China Suzhou Laboratory Suzhou China

ISBN: (纸本)9798891761148

Recently, Large Language Models (LLMs) have been demonstrated to possess impressive capabilities in a variety of domains and tasks. We investigate the issue of prompt design in the multi-turn text-to-SQL task and attempt to enhance the LLMs’ reasoning capacity when generating SQL queries. In the conversational context, the current SQL query can be modified from the preceding SQL query with only a few operations due to the context dependency. We introduce our method called CoE-SQL1 which can prompt LLMs to generate the SQL query based on the previously generated SQL query with an edition chain. We also conduct extensive ablation studies to determine the optimal configuration of our approach. Our approach outperforms different in-context learning baselines stably and achieves state-of-the-art performances on two benchmarks SParC and CoSQL using LLMs, which is also competitive to the SOTA fine-tuned models. © 2024 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

A novel deep learning model for early diabetes risk prediction using attention-enhanced deep belief networks with highly imbalanced data

引用

International Journal of Information Technology (Singapore) 2025年第4期17卷 1933-1955页

作者： Olabanjo, Olusola Wusu, Ashiribo Olabanjo, Olufemi Asokere, Mauton Afisi, Oseni Akinnuwesi, Boluwaji Center for Equitable AI and Machine Learning Systems (CEAMLS) Morgan State University Baltimore 21251 MD United States Department of Mathematics Morgan State University Baltimore MD United States Department of Computer Science Lagos State University Lagos Nigeria Department of Mathematics Lagos State University Lagos Nigeria Department of Cybersecurity University of Derby Derby United Kingdom Department of Philosophy Lagos State University Lagos Nigeria Department of Computer Science Faculty of Science and Engineering University of Eswatini Kwaluseni Swaziland

Diabetes mellitus is a prevalent chronic illness with severe complications that demand timely diagnosis. This study introduces an attention-enhanced Deep Belief Network (DBN) for early diabetes risk prediction, designed to address challenges associated with highly imbalanced datasets. Using a dataset from Sylhet Diabetes Hospital, which includes symptom and demographic information from patients, we applied an ensemble feature selection approach to identify critical predictors. To address the class imbalance, Generative Adversarial Networks (GANs) were used to generate synthetic data, ensuring the model’s robustness in identifying underrepresented cases. Additionally, a hybrid loss function combining cross-entropy and focal loss was implemented to improve classification, especially for hard-to-detect instances. Our results show that the attention-based DBN model, augmented with synthetic data from GANs and optimized with a hybrid loss function, achieves an AUC of 1.00, F1-score of 0.97, precision of 0.98, and recall of 0.95, outperforming several baseline models. This research offers a novel and effective approach for early diabetes detection, demonstrating potential for use as a clinical tool in preventive healthcare settings. © The Author(s) 2025.

关键词： Attention mechanism Deep belief networks (DBN) Diabetes risk Generative adversarial networks (GANs) Hybrid loss function Imbalanced data

来源：评论

学校读者我要写书评

暂无评论

Li-Fi Enables Reliable Communication of VLT for Secured Data Exchange

Li-Fi Enables Reliable Communication of VLT for Secured Data...

引用

International Conference on Intelligent Systems and Sustainable Computing, ICISSC 2021

作者： Gupta, N. Venkata Ramana Rajeshkumar, G. Selvi, S. Arun Mozhi Sengan, Sudhakar Subburaj, Arjun Priyadarsini, S. Rangasamy, Rajasekar Department of Computer Science and Engineering Prasad V. Potluri Siddhartha Institute of Technology Vijayawada Kanuru India Department Computer Science and Engineering Erode Sengunthar Engineering College Tamil Nadu Perundurai Erode India Department of Computer Science and Engineering DMI St. John The Baptist University Mangochi Malawi Department of Computer Science and Engineering PSN College of Engineering and Technology Tamil Nadu Tirunelveli India Tranxit Technology Solutions Pvt. Ltd. Tamil Nadu Chennai India Department of Computer Science and Engineering P.S.R Engineering College Tamil Nadu Sivakasi India Department of Computer Science and Engineering-AI & ML School Engineering Malla Reddy University Hyderabad India

ISBN: (纸本)9789811900105

Li-Fi is a new method of sending and receiving secure data using light energy from one node to another. When data are transferred to the receiver node, the sender’s node light will blink continuously in the decade research. This may appear to be an uncomfortable remote control. Other than that, there are disadvantages to using a Wi-Fi network for the node because it has significant drawbacks such as bandwidth variances between nodes when using the communications system, such as some nodes achieving actual transmission speeds and others having low data rates. Li-Fi solves this issue by using light energy similar on both the sender and receiver sides, so the receiver node can transfer data at high speeds even when the device sending it is far away. The data transfer speed will be 9600 bits per second, which is fast enough to prevent lagging. The Li-Fi system has mainly been used in wireless nodes and IoT devices with embedded technology, a colour sensor and a microcontroller. The virtual node’s light source communicates with the destination nodes’ colour sensor, transforming the light into digital information. For the most part, these apps are found in IoT devices for user needs. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

DCBLCCG NET: A Fusion Approach For Enhanced Pepper Leaf Disease Identification With Hybrid Deep Learning

DCBLCCG NET: A Fusion Approach For Enhanced Pepper Leaf Dise...

引用

Intelligent Systems and Advanced Applications (ICISAA), International Conference on

作者： Shaik Salma Asiya Begum Hussainbi Shaik Hussain Syed Department of CSE(AI&ML) LakiReddy Bali Reddy College of Engineering Mylavaram NTR District AP India School of Computer Science and Engineering VIT-AP University Amaravati AP India

ISBN: (数字)9798331539948

ISBN: (纸本)9798331539955

The agricultural sector in developing countries like Ethiopia heavily relies on crop health and productivity. Pepper, a crucial crop for global food security, faces significant threats from various diseases, including blight leaf disease, gray leaf spot, common rust, fruit rot, and powdery mildew. The existence of 34 different pepper diseases has resulted in an average reduction of 33% in crop productivity. Traditional disease detection methods, such as visual inspection by farmers, are prone to errors and are time-consuming Previous research has largely employed image processing and deep learning for disease classification in pepper plants. This study introduces a novel approach by combining a Deep Convolutional Neural Network (DCNN) with a Bidirectional Long Short-Term Memory (BILSTM) network. While DCNN excels at extracting spatial features, BILSTM captures the sequential growth patterns of pepper plants. This hybrid model achieved $\mathbf{1 0 0 \%}$ accuracy during training and performed exceptionally well on a dataset of pepper disease images. Utilizing features from pretrained CNNs like VGG16 and AlexNet, the model attained a top classification accuracy of $\mathbf{9 9. 5 \%}$, with grayscale datasets and segmented images achieving $\mathbf{9 8. 9 \%}$ and $\mathbf{9 9. 3} \%$, respectively. This robust model is a powerful tool for rapid and accurate pepper disease identification, improving disease management and agricultural productivity.

关键词： Productivity Image segmentation Accuracy Microorganisms Transfer learning Bidirectional long short term memory Crops Gray-scale Feature extraction Diseases

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：