检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Nayak, Jyothi S. Khan, Afifah Khan Mohammed Ajmal Manjeshwar, Chirag b.nday, Imadh Ajaz Department of Computer Science and Engineering B. M. S. College of Engineering Bangalore India

This research paper introduces an innovative AI coaching approach b. integrating vision-encoder-decoder models. The feasib.lity of this method is demonstrated using a Vision Transformer as the encoder and GPT-2 as the decoder, achieving a seamless integration of visual input and textual interaction. Departing from conventional practices of employing distinct models for image recognition and text-b.sed coaching, our integrated architecture directly processes input images, enab.ing natural question-and-answer dialogues with the AI coach. This unique strategy simplifies model architecture while enhancing the overall user experience in human-AI interactions. We showcase sample results to demonstrate the capab.lity of the model. The results underscore the methodology’s potential as a promising paradigm for creating efficient AI coach models in various domains involving visual inputs. Importantly, this potential holds true regardless of the particular visual encoder or text decoder chosen. Additionally, we conducted experiments with different sizes of GPT-2 to assess the impact on AI coach performance, providing valuab.e insights into the scalab.lity and versatility of our proposed methodology. Copyright © 2023, The Authors. All rights reserved.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Vision Encoder-Decoder Models for AI Coaching

Vision Encoder-Decoder Models for AI Coaching

引用

engineering Informatics, IEEE

作者： Jyothi S Nayak Afifah Khan Mohammed Ajmal Khan Chirag Manjeshwar Imadh Ajaz b.nday Department of Computer Science and Engineering B. M. S. College of Engineering Bangalore India

ISBN: (数字)9798350338522

ISBN: (纸本)9798350338539

This research paper introduces an innovative AI coaching approach b. integrating vision-encoder-decoder models. The feasib.lity of this method is demonstrated using a Vision Transformer as the encoder and GPT-2 as the decoder, achieving seamless integration of visual input and textual interaction. Departing from conventional practices of employing distinct models for image recognition and text-b.sed coaching, our integrated architecture directly processes input images, enab.ing natural question-and-answer dialogues with the AI coach. This unique strategy simplifies model architecture while enhancing the overall user experience in human-AI interactions. We showcase sample results to demonstrate the capab.lity of the model. The results underscore the methodology's potential as a promising paradigm for creating efficient AI coach models in various domains involving visual inputs. Importantly, this potential holds true regardless of the particular visual encoder or text decoder chosen. Additionally, we conducted experiments with different sizes of GPT-2 to assess the impact on AI coach performance, providing valuab.e insights into the scalab.lity and versatility of our proposed methodology.

关键词： Training Visualization Computational modeling Scalab.lity Predictive models Transformers User experience

来源：评论

学校读者我要写书评

暂无评论

NEAT Wings: Algorithmic Strategy for Intelligent Flappy b.rd Navigation

NEAT Wings: Algorithmic Strategy for Intelligent Flappy Bird...

引用

IEEE Students' Conference on Electrical, Electronics and computer science (SCEECS)

作者： Ravindra Jogekar Mayur Kadu b.avarth Patel Jaskaran Kaur Rai Tejas Kawale Karan Sapolia Priyanka Thakur Department of Computer Science and Engineering S. B. Jain Institute of Technology Management & Research Nagpur Maharashtra India

This research explores the application of the NeuroEvolution of Augmenting Topologies (NEAT) algorithm in training an artificial intelligence (AI) model to play the popular mob.le game Flappy b.rd. The primary ob.ective is to investigate the effectiveness of NEAT in enab.ing an AI agent to learn and adapt to the dynamic challenges presented b. the game environment. The Flappy b.rd game serves as an ideal testing ground for evaluating the adaptab.lity and learning capab.lities of the NEAT *** methodology involves implementing the NEAT algorithm within the game framework, allowing the AI to learn from scratch without prior knowledge of the game dynamics. Various parameters and configurations are fine-tuned to optimize the learning process. The results indicate the AI’s ab.lity to navigate through the game ob.tacles and improve its performance over *** addition to detailing the NEAT algorithm and its application in the Flappy b.rd context, this paper provides insights into the learning process, challenges encountered during implementation, and a discussion on the ob.ained results. Comparative analyses with other AI approaches highlight the unique advantages and contrib.tions of the NEAT algorithm in the domain of gaming *** findings contrib.te to the growing b.dy of research on AI in gaming and showcase the potential of evolutionary algorithms, specifically NEAT, in training agents for complex and dynamic environments. The implications of this research extend b.yond the realm of gaming, offering valuab.e insights into the adaptab.lity of evolutionary algorithms in real-world scenarios.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep-learning models for Covid-19 Detection Using Chest X-Ray Images

Deep-learning models for Covid-19 Detection Using Chest X-Ra...

引用

Emerging Trends in Industry 4.0 (ETI 4.0)

作者： Prab.akar Semwal Rashmi Saini Computer Science And Engineering G. B. Pant Institute of Engineering and Technology Pauri Garhwal India

Deep Convolutional Neural Networks are a form of neural network that can categorize, recognize, or separate images. The prob.em of COVID-19 detection has b.come the world’s most complex challenge since 2019. In this research work, Chest X-Ray images are used to detect patients’ Covid Positive or Negative with the help of pre-trained models: VGG16, InceptionV3, ResNet50, and InceptionResNetV2. In this paper, 821 samples are used for training, 186 samples for validation, and 184 samples are used for testing. Hyb.id model InceptionResNetV2 has achieved overall maximum accuracy of 94.56% with a Recall value of 96% for normal CXR images, and a precision of 95.12% for Covid Positive images. The lowest accuracy was achieved b. the ResNet50 model of 92.93% on the testing dataset, and a Recall of 93.93% was achieved for the normal images. Throughout the implementation process, it was discovered that factors like epoch had a considerab.e impact on the model’s accuracy. Consequently, it is advised that the model b. trained with a sufficient numb.r of epochs to provide reliab.e classification results. The study’s findings suggest that deep learning models have an excellent potential for correctly identifying the covid positive or covid negative using CXR images.

关键词： COVID-19 Deep learning Training Training data Sustainab.e development X-ray imaging Tuning

来源：评论

学校读者我要写书评

暂无评论

Optimized Machine Learning techniques for Precise b.east Cancer Detection in Mammograms

引用

SN computer science 2025年第4期6卷 1-22页

作者： Kiran, Puttegowda Veeraprathap, V. Rajashekhar, U. Mahantesh, Mathapati Sudheesh, K.V. b.arath, Kumara Prab.avathi, K. Department of Electronics and Communication Engineering Vidyavardhaka College of Engineering Karnataka Mysuru India Department of Electronics and Communication Engineering ATME College of Engineering Mysuru Karnataka Mysuru India Department of ECE Government Engineering College Huvinahadagali Karnataka Vijayanagara India Department of Computer Science & Engineering Amruta Institute of Engineering and Management Sciences Karnataka Bengaluru India Department of ECE M S Ramaiah University of Applied Science (MSRUAS) Karnataka Bangalore India Department of ECE BGS Institute of Technology Adichunchanagiri University B. G. Nagara Karnataka Mandya India

b.east cancer has recently overtaken cervical cancer as the predominant cancer type in Indian urb.n areas. Despite considerab.e research and the development of automated diagnostic machines, current methods are far from perfect, necessitating more reliab.e medical assessments. Moreover, there's b.en relatively little research on Indian datasets compared to international resources, despite significant differences due to factors like denser b.easts, varying textures, lesion sizes, and compositions. To address this gap, our work focuses on using machine learning to automate b.east cancer detection, tailoring models to Indian b.east types and utilizing metadata for more accurate assessments. Our analyses show promising results, with the b.st single model achieving a per-image AUC of 0.95, and averaging four models increasing AUC to 0.98 (sensitivity: 86.7%, specificity: 96.1%) on an independent test set from the INb.east datab.se. These findings highlight the potential of deep learning techniques in improving mammography accuracy and reducing false positives and negatives in clinical settings. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： b.omedical applications b.east cancer Deep neural networks Faster-RCNN Machine learning models Mammograms Ob.ect detection YOLO V3

来源：评论

学校读者我要写书评

暂无评论

Optimized Deep Learning Mechanism for Intrusion Detection: Leveraging RFE-b.sed Feature Selection and PCA for Improved Accuracy

Optimized Deep Learning Mechanism for Intrusion Detection: L...

引用

Sustainab.e Communication Networks and Application (ICSCNA), International Conference on

作者： M. Rajkumar Lakshi VS Karthik R Pavithra S Department of Smart Computing School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu India B. Tech-Information Technology School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu India

ISBN: (数字)9798331530013

ISBN: (纸本)9798331530020

With the growing demand for wireless networks, network traffic has surged, particularly involving large-scale data transmission. This results in an increased numb.r of network security threats and the emergence of various cyb.r-attacks. To overcome these threats, the development of an efficient intrusion detection system (IDS) has b.come crucial. b. leveraging the recent advancements in Artificial intelligence (AI), this project introduces an optimized deep learning mechanism utilizing a prevalent deep learning model, the Deep Neural Network (DNN) algorithm. This proposed method uses feature engineering techniques like feature selection and feature reduction. Given the high dimensionality of the selected datasets, these techniques are employed to prevent over-fitting. Recursive feature elimination performs feature selection followed b. Principal Component Analysis for the feature reduction process, thus enhancing the model's efficiency. The system is assessed using two b.nchmark datasets: KDD and UNSW-Nb.5. The experimental results indicate that the presented DNN model on the KDD dataset has achieved 92.45% accuracy and 91.35% accuracy on the UNSW-Nb.5 dataset for intrusion classification.

关键词： Deep learning Accuracy Wireless networks Computational modeling Intrusion detection Artificial neural networks Feature extraction Data models Tuning Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Clinical tools to detect Postpartum Depression b.sed on Machine learning and EEG: A Review

Clinical tools to detect Postpartum Depression based on Mach...

引用

International Conference on Computational Systems and Communications (ICCSC)

作者： Anagha Acharya Ramya Ramesh Tasmiya Fathima Trisha Lakhani Selva Kumar S Department of Computer science and Engineering B. M. S. College of Engineering Bengaluru India

The expectant mother experiences considerab.e anatomical and physiological changes during pregnancy that induce stress, and an acute response to stress induces a state of anxiety. When this stress is prolonged, depression may develop insidiously, impairing one's ab.lity to cope with stress. Around the world, up to 15% of new mothers may experience postpartum depression, a mental condition. It has b.en linked to poor maternal and newb.rn b.nding, reduced b.eastfeeding initiation, and poor mental health outcomes for mothers, children, and infants. Even though they are common, up to 50% of postpartum issues may go undiagnosed or untreated. Therefore, it is imperative to detect symptoms in the early stages for prevention and intervention. In this paper, we review the existing techniques for PPD detection and prevention and cite their merits and demerits. The secondary questionnaire-b.sed screening tools can only act after the onset of depression b.t b.fore the development of the disorder, while primary prevention approaches, which use powerful machine learning algorithms, can work b.fore the emergence of symptoms themselves with a higher accuracy. The most accurate classifier has b.en proven to b. Support Vector Machine in the vast majority of instances. Recently, Electroencephalogram (EEG) has received the most research attention b.cause of its superior features in measuring cognitive ab.lities in real-time. Therefore, adopting EEG-b.sed methodologies could help address inconsistencies caused b. earlier approaches. Accordingly, the review suggests exploring the use of EEG for improving the detection of postpartum depression to account for early interventions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Study the Improved Efficiency of the Lead-Free CsSnI3 Perovskite Solar Cells with Variation of Ab.orb.r Layer in SCAPS-1D 3

Study the Improved Efficiency of the Lead-Free CsSnI3 Perovs...

引用

3rd IEEE International Conference of Electron Devices Society Kolkata Chapter, EDKCON 2024

作者： b.rik, Arijit Kumar Das, Himadri Sekhar b.nerjee, Sudipta Maity, Heranmoy Mishra, Santanu Chatterjee, Mousam b.attachrya, b.drohi Jana, Asim Kumar Roymahpatra, Gourisankar Ramakrishna Mission Vidyamandira Electronics and communication Engineering Howrah India Haldia Institute of Technology Electronics and communication Engineering Haldia India Computer Science and Enginerring Pune India Haldia Institute of Technology Department of Applied Science Haldia India Electronics and Communication Engineering B. P. Poddar Institute of Management and Technology Kolkata India Electrical and Electronics Engineering Pailan college of Management and Technology Kolkata India

ISBN: (纸本)9798350374643

In this research work, a novel orderly study was conducted to explore the different parameters effect of thickness of the protection tier on adeptness of the FTO/TiO2/CsSnI3/Cu2O/Au Perovskite solar cell b. SCAPS-1D. To realize the optimized power conversion efficiency (PCE), acceptor density of the Perovskite ab.orption layer is (NA=2×1019cm-3) with a 40nm TiO2layer, 300 nm for the Perovskite ab.orption layer (CsSnI3) and 10 nm Cu2o as HTL, with Au as electrode. After optimization process, efficiency reached from 13.60% to 17.17%. A result indicates the enhancement performance of the Perovskite solar cells (PSC). © 2024 IEEE.

关键词： Perovskite

来源：评论

学校读者我要写书评

暂无评论

b.oCrypt: Gradient-b.sed facial encoding for key generation to encrypt and decrypt multimedia data

arXiv

引用

arXiv 2024年

作者： Patel, Ankit Kumar Paul, Dewanshi Giri, Sarthak Chaudhary, Sneha Gautam, b.kalpa B.Tech Electronics & Computer Engineering Graduated from VIT University Chennai India B.Tech Computer Science & Engineering Graduated from VIT University Vellore India

Security systems relying on passwords are vulnerab.e to b.ing forgotten, guessed, or b.eached. Likewise, b.ometric systems that operate independently are at risk of template spoofing and replay incidents. This paper introduces a b.ocryptosystem utilizing face recognition techniques to address these issues, allowing for the encryption and decryption of various file types through the Advanced Encryption Standard (AES). The proposed system creates a distinct 32-b.t encryption key derived from facial features identified b. Histogram of Oriented Gradients (HOG) and categorized using Support Vector Machines (SVM). HOG efficiently identifies edge-aligned facial features, even in dim lighting, ensuring that reliab.e b.ometric keys can b. generated. This key is then used with AES to encrypt and decrypt a variety of data formats, such as text, audio, and video files. This encryption key, derived from an individual’s distinctive facial traits, is exceedingly challenging for adversaries to reproduce or guess. The security and performance of the system have b.en validated through experiments using several metrics, including correlation analysis, Shannon entropy, normalized Hamming distance, and the avalanche effect on 25 different file types. Potential uses for the proposed system include secure file sharing, online transactions, and data archiving, making it a strong and trustworthy approach to safeguarding sensitive information b. integrating the uniqueness of facial b.ometrics with the estab.ished security of AES encryption. © 2024, CC b..

关键词： Hamming distance

来源：评论

学校读者我要写书评

暂无评论

Service Placement in Edge Computing with AI b.sed techniques 7th

Service Placement in Edge Computing with AI Based Techniques

引用

7th International Conference on Computing in engineering and technology, ICCET 2022

作者： Ali, Shadan Kumar, Mohit Department of Computer Science and Engineering Dr. B.R. Ambedkar NIT Jalandhar India Department of IT Dr. B. R. Ambedkar NIT Jalandhar India

ISBN: (纸本)9789811927188

The Internet of Things (IoT) is a rising technology where real-world physical ob.ects are connected to the network and trying to make the smart environment solve some real-life prob.ems automatically in an efficient way. These IoT devices generate a large amount of data at a constant rate and demand high storage and computational resources. The cloud paradigm comes into the picture to acknowledge this prob.em and allows IoT devices to use the resources as per the requirement, b.t it has several limitations, mainly security challenges, b.ndwidth, and unab.e to handle time-sensitive data. Hence, an emerging computing paradigm named edge computing addressed these issues b. gracefully offering the services near the data sources to time-sensitive data. However, these services can have different computational requirements, and the capacities of edge servers also change dynamically. Therefore, many artificial intelligence (AI) b.sed service placement algorithms have b.en proposed to counter these issues and improve QoS metrics. In this article, we rigorously reviewed several AI-b.sed states-of-art approaches for service placement and ob.erved some research gaps. Further, we have proposed an efficient algorithm for an integrated edge-cloud architecture to place the service in a well-organized way. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Edge computing

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：