检索结果-内蒙古大学图书馆

Layer-Wise Learning Rate Optimization for Task-Dependent Fine-Tuning of Pre-Trained Models: An Evolutionary Approach

引用

ACM Transactions on Evolutionary Learning and Optimization 2024年第4期4卷 1-23页

作者： Bu, Chenyang Liu, Yuxin Huang, Manzong Shao, Jianxuan Ji, Shengwei Luo, Wenjian Wu, Xindong Key Laboratory of Knowledge Engineering with Big Data Ministry of Education and School of Computer Science and Information Engineering Hefei University of Technology Hefei China School of Artificial Intelligence and Big Data Hefei University Hefei China Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies School of Computer Science and Technology Harbin Institute of Technology Shenzhen China

The superior performance of large-scale pre-Trained models, such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-Trained Transformer (GPT), has received increasing attention in both academic and industrial research and has become one of the current research hotspots. A pre-Trained model refers to a model trained on large-scale unlabeled data, whose purpose is to learn general language representation or features for fine-Tuning or transfer learning in subsequent tasks. After pre-Training is complete, a small amount of labeled data can be used to fine-Tune the model for a specific task or domain. This two-stage method of "pre-Training+fine-Tuning"has achieved advanced results in natural language processing (NLP) tasks. Despite widespread adoption, existing fixed fine-Tuning schemes that adapt well to one NLP task may perform inconsistently on other NLP tasks given that different tasks have different latent semantic structures. In this article, we explore the effectiveness of automatic fine-Tuning pattern search for layer-wise learning rates from an evolutionary optimization perspective. Our goal is to use evolutionary algorithms to search for better task-dependent fine-Tuning patterns for specific NLP tasks than typical fixed fine-Tuning patterns. Experimental results on two real-world language benchmarks and three advanced pre-Training language models show the effectiveness and generality of the proposed framework. © 2024 held by the owner/author(s).

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

SafeRoute: Pioneering Online Ride Booking with Pega & ML-Powered Safety Measures 8

SafeRoute: Pioneering Online Ride Booking with Pega & ML-Pow...

引用

8th International Conference on Electronics, Communication and Aerospace Technology, ICECA 2024

作者： Rajalakshmi, D. Immaculate, P.S. Pandi, S. Senthil Pavitra, A. Rehash Rushmi Vijayalakshmi, V. Sharanya, S. Srm Institute of Science and Technology Data Science and Business Systems Chennai India Sri Sai Ram Institute of Technology Chennai Master of Business Administration India Rajalakshmi Engineering College Computer Science and Engineering Chennai India

ISBN: (纸本)9798350367904

The advent of online ride booking services has revolutionized the transportation industry, providing users with convenient and efficient means of commuting. However, ensuring the safety of passengers and drivers remains a paramount concern. This project introduces an innovative solution that leverages Pega, a leading business process management platform, coupled with machine learning algorithms to enhance the safety measures within an online ride booking service. The proposed system integrates seamlessly into the existing online ride booking infrastructure and introduces advanced safety mechanisms based on realtime data analysis. Machine learning models are employed to predict and mitigate potential risks, ensuring a secure and reliable experience for both passengers and drivers. © 2024 IEEE.

关键词： Reservation systems

来源：评论

学校读者我要写书评

暂无评论

Investigation of Handwritten Image-To-Speech Using Deep Learning

Investigation of Handwritten Image-To-Speech Using Deep Lear...

引用

2024 International Conference on Advances in Modern Age Technologies for Health and engineering science, AMATHE 2024

作者： Manju, S. Anitha, J. Sujitha Juliet, D. Karunya Institute Of Technology And Sciences Division Of Computer Science And Engineering Coimbatore India Karunya Institute Of Technology And Sciences Division Of Data Science And Cyber Security Coimbatore India

ISBN: (数字)9798350371567

ISBN: (纸本)9798350371567

In the realm of accessibility technology, this paper introduces a pioneering method for converting handwritten images to speech. The work primarily focuses on recognizing handwritten text and subsequently converting it into audible speech. The work comprises two fundamental phases: initially, a deep learning algorithm is employed to recognize and extract text from handwritten images precisely. Following this, the recognized text is converted into speech using the pyttsx3 library which offers flexibility in voice selection, speech rate, and other parameters, making it highly customizable for different applications. To ascertain the most effective algorithm for the text recognition phase, a comparative analysis is conducted among the top three algorithms-Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Transformer-based models - based on their accuracy in recognizing handwritten text. Experimental results indicate that CNNs surpass the other two algorithms, achieving an accuracy of 92%. This work offers a pragmatic approach to converting handwritten content into speech, with potential applications in aiding the visually impaired and enhancing accessibility to written information. © 2024 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

VGG-16 based Deep Learning Approach for Plant Species Detection

VGG-16 based Deep Learning Approach for Plant Species Detect...

引用

2024 International Conference on Advances in Modern Age Technologies for Health and engineering science, AMATHE 2024

作者： Praveen, Bandi Isravel, Deva Priya Malar Dhas, Julia Punitha Karunya Institute Of Technology And Sciences Division Of Data Science And Cyber Security Coimbatore India Karunya Institute Of Technology And Sciences Division Of Computer Science And Engineering Coimbatore India

ISBN: (纸本)9798350371567

Automated plant species detection has gained interest due to the potential benefits in agriculture to significantly increase crop yields, decrease the use of chemicals, and promote sustainable agriculture practices. A deep learning model, for plant species identification is a promising approach. This paper presents a condensed version of the VGG-16 deep learning model for identifying plant species. The original VGG-16 model's complexity and computational overhead are minimized by lowering the number of convolutional layers, which lowers the amount of parameters and calculations needed while keeping a high level of accuracy in the identification of various plant species. The streamlined VGG-16 model obtains an improved accuracy rate in categorizing distinct plant species after being trained on a big dataset of plant species that includes a variety of images of different plants. The condensed VGG-16 model offers a viable method for precise and effective plant species identification. This study's methodology and results advances the area of plant species detection by offering a deeper learning model that is more effective and useful and can be used on devices with limited resources. © 2024 IEEE.

关键词： Agriculture

来源：评论

学校读者我要写书评

暂无评论

A Deepfake Image Classifier System for real and Doctored Image Differentiation 7

A Deepfake Image Classifier System for real and Doctored Ima...

引用

7th International Conference on Inventive Computation Technologies, ICICT 2024

作者： Thomas, Juanith Mathew Ebenezer, V. Richard, Rohan Paul Karunya Institute of Technology And Sciences Division of Computer Science And Engineering Coimbatore India Karunya Institute of Technology And Sciences Division of Data Science And Cyber Security Coimbatore India

ISBN: (纸本)9798350359299

The sudden proliferation of deepfake technology has raised concerns about the authenticity and integrity of digital media. In response to these concerns, this paper presents a Deepfake image classifier system designed to differentiate between real and doctored images using "MesoNet"architecture which utilizes convolutional layers and builds a suitable neural network for classification. Making efficient use of Deep Learning strategies, this system aims to detect subtle manipulations indicative of deepfake alterations. The proposed model is trained on two distinct datasets: dataset I comprises authentic images while dataset II consists of images subjected to various levels of manipulation, making them deepfakes. Through extensive training using these images and their labels, we'll train the neural network to adequately predict the authenticity of an image. The model has been trained on exactly 7104 images and judging from the results, the predictions made by the model are fairly accurate. Results yield an accuracy of 88.81% and a precision of 87.93%. © 2024 IEEE.

关键词： Digital storage

来源：评论

学校读者我要写书评

暂无评论

Ontological Knowledge-Based Representation System to Prevent Web Attacks in iMANET's

Ontological Knowledge-Based Representation System to Prevent...

引用

2024 International Conference on Communication, Computer sciences and engineering, IC3SE 2024

作者： Chandani, Priyanka Gupta, Anand Kumar Manivasagam Noida Institute Of Engineering And Technology Department Of Data Science Uttar Pradesh Greater Noida India Tula's Institute Department Of Computer Science And Engineering Dehradun India Department Of Computer Science And Information Technology Bangalore India

ISBN: (纸本)9798350366846

This paper implies securing Internet-based Mobile Ad-hoc Networks (iMANETs) with semantic web techniques. To illustrate safety issues, ontologies will be utilised rather than taxonomies. These ontologies can be added to an intrusion detection system's (IDS) information repository. By employing inference based on semantic relations, intrusion detection systems can identify and incorporate upcoming attacks into their current knowledge base. © 2024 IEEE.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

Criminal Identification Based On database Management Using Blockchain 4

Criminal Identification Based On Database Management Using B...

引用

4th International Conference on Power, Energy, Control and Transmission Systems, ICPECTS 2024

作者： Suganya, R. Ashok, P. Lekshitha, J. Shanmuga Priya, S. Monica, K. Somasundaram, K. Sri Sai Ram Institute of Technology Department of Computer Science and Engineering Chennai India Saveetha Engineering College Department of Artificial Intelligence and Data Science Chennai India Sri Muthukumaran Institute of Technology Department of Computer Science and Engineering Chennai India

ISBN: (数字)9798331508845

ISBN: (纸本)9798331508845

The crime monitoring system is a unique and authentic project which functions with the concepts of block chain language. Blockchain technology has the potential to revolutionize the management of criminal records by providing a secure network where documents are readily accessible and immune to tampering. This innovation ensures the safety and integrity of sensitive information, offering a robust solution to the challenges posed by traditional recordkeeping system. The basic use of the crime identification system is to maintain and process the criminal records and data that are given by the user. This can be used in the security system during the arrival of dignitary people for the occasional events by recognizing the people based upon the criminal records. Thus, criminals are restricted from entering inside the event. This can also be embedded for the security cameras of the event and are constantly monitored. © 2024 IEEE.

关键词： Records management

来源：评论

学校读者我要写书评

暂无评论

Attention based 1-D CNN system for detection of Freezing of gait (FOG) in Parkinsons Patients 15

Attention based 1-D CNN system for detection of Freezing of ...

引用

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

作者： Pradeep Kumar, D. Sowmya, B.J. Vincent, Ashvin Pandiyan, Aravindh Ramaiah Institute of Technology Department of Computer Science and Engineering Bangalore India Ramaiah Institute of Technology Department of Artificial Intelligence and Data Science Bangalore India Bangalore India

ISBN: (纸本)9798350370249

One common clinical symptom seen in Parkinson's disease (PD) patients is freezing of gait (FOG). It manifests as an irregular gait, marked by abrupt, involuntary stopping of movement during gait episodes. FOG entails a major negative effect on PD patients' quality of life. In this work, we analyze the movement data that is collected using the lower-back 3D accelerometer worn by the patients during FOG episodes. By employing machine learning methods to classify this data, we aim to accurately detect FOG trends. In particular, we focus on Deep Learning-based techniques, specifically an attention-based 1D Convolutional Neural Network (1D-CNN), due to their capacity to extract intricate features from unprocessed sensor input. Our results indicate that the proposed model achieves AUC scores of 0.9032 with respect to the TDCSFOG dataset and 0.8682 for the DeFOG dataset, for a combined average AUC of 0.8857. Mean Average Precision (mAP) scores for TDCSFOG and DeFOG reach 0.8388 and 0.779, respectively, for an average of 0.8089, showing the ability of the model to predict FOG occurrences over time. These findings suggest techniques based on Deep Learning enhance FOG detection, allowing clinicians to more effectively tailor treatments and care strategies for Parkinson's patients, thus enhancing their ability to move and overall well-being. © 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

Ensemble Machine Learning based Weather Prediction System 2

Ensemble Machine Learning based Weather Prediction System

引用

2nd IEEE International Conference on Communication, Security and Artificial Intelligence, ICCSAI 2023

作者： Vijayaganth, V. Monicaa Jayasri, P.K. Narmadha, S. Pavithra Devi, S. Kpr Institute of Engineering and Technology Department of Artificial Intelligence and Data Science Coimbatore India

ISBN: (纸本)9798350369960

The technique of predicting the weather using science and technology for a particular area is known as weather forecasting, we employed the ensemble approach to produce more precise results. Ensemble learning enhances machine learning outcomes by mixing numerous models. The predictive performance of this strategy is superior to that of a single model. Many machine learning methods and algorithms can anticipate changes in the weather. According to prior studies, there are many additional weather forecasting methods. Depending on various factors like temperature, wind, humidity, and rainfall. © 2023 IEEE.

关键词： Ensemble learning Forecasting Methods k-NN Machine Learning Rainfall SVM Weather XG Boost

来源：评论

学校读者我要写书评

暂无评论

Exploring Novel Machine Learning Approaches for Cheat Detection in Competitive Gaming 5

Exploring Novel Machine Learning Approaches for Cheat Detect...

引用

5th International Conference on Information Management and Machine Intelligence, ICIMMI 2023

作者： Singh, Uday Pratap Kumari, Bersha Goyal, Dinesh Afrid, Mohd Department of Artificial Intelligence & Data Science Poornima Institute of Engineering & Technology Jaipur India

ISBN: (纸本)9798400709418

The issue of cheating in competitive gaming is a major concern that undermines fair play and the overall gaming experience. Although machine learning has proven to be an effective tool for detecting cheats, there is still untapped potential for innovative approaches that can further improve anti-cheat systems. This research paper delves into unexplored machine learning methods that hold promise for future cheat detection strategies. These include Graph Neural Networks (GNNs) for multifaceted analysis, reinforcement learning-based cheat adaptation, federated learning for privacy-preserving cheat detection, Explainable AI (XAI) for transparent cheat detection, Generative Adversarial Networks (GANs) for anomaly detection, unsupervised learning for unlabeled data analysis, active learning for efficient data collection, continuous learning for evolving cheat landscapes, multi-agent reinforcement learning for collaborative cheat detection, and transfer learning for cross-game cheat detection. By leveraging these novel techniques, the gaming industry can develop more robust and comprehensive cheat detection solutions that uphold the integrity of competitive gaming environments. © 2023 ACM.

关键词： Unsupervised learning

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：