检索结果-内蒙古大学图书馆

Classification and study of music genres with multimodal Spectro-Lyrical Embeddings for Music (SLEM)

Multimedia Tools and Applications 2025年第7期84卷 3701-3721页

作者： Mehra, Ashman Mehra, Aryan Narang, Pratik Department of Computer Science and Information Systems Birla Institute of Technology and Science Goa Pilani India Computer Science Department Carnegie Mellon University PittsburghPA United States Department of Computer Science and Information Systems Birla Institute of Technology and Science Pilani Rajasthan Pilani India

The essence of music is inherently multi-modal – with audio and lyrics going hand in hand. However, there is very less research done to study the intricacies of the multi-modal nature of music, and its relation with genres. Our work uses this multi-modality to present spectro-lyrical embeddings for music representation (SLEM), leveraging the power of open-sourced, lightweight, and state-of-the-art deep learning vision and language models to encode songs. This work summarises extensive experimentation with over 20 deep learning-based music embeddings of a self-curated and hand-labeled multi-lingual dataset of 226 recent songs spread over 5 genres. Our aim is to study the effects of varying the weight of lyrics and spectrograms in the embeddings on the multi-class genre classification. The purpose of this study is to prove that a simple linear combination of both modalities is better than either modality alone. Our methods achieve an accuracy ranging between 81.08% to 98.60% for different genres, by using the K-nearest neighbors algorithm on the multimodal embeddings. We successfully study the intricacies of genres in this representational space, including their misclassification, visual clustering with EM-GMM, and the domain-specific meaning of the multi-modal weight for each genre with respect to ’instrumentalness’ and ’energy’ metadata. SLEM presents one of the first works on an end-to-end method that uses spectro-lyrical embeddings without hand-engineered features. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Music

来源：评论

学校读者我要写书评

暂无评论

Satellite Image to Map Translation using GANs 15

Satellite Image to Map Translation using GANs

引用

15th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2024

作者： Phatangare, Sheetal Khalifa, Mohammed Mansoor Kharche, Sejal Khatib, Amaan Kshirsagar, Atharva Vishwakarma Institute of Technology Computer Engineering Department Pune India

ISBN: (纸本)9798331300579

Generative adversarial networks (GANs) are a promising method for learning deep representations without spending a lot of time on data for annotated training. They accomplish this by obtaining signals that are back propagated through a competitive mechanism in-transforming two networks. The possible representations of a variety of applications may employ the knowledge gathered by GANs, image synthesis, semantic image editing, fashion categorization, transmission, and image super-resolution. The goal of this research article is to give an overview of GANs. the signal processing community, utilizing well-known whenever feasible, use concepts and analogies. Additionally, to recognize various training techniques and building GANs, we also highlight lingering issues with their theory and utilization. © Grenze Scientific Society, 2024.

关键词： Gallium nitride

来源：评论

学校读者我要写书评

暂无评论

Forecasting Renewable Energy Generation: A Predictive Modelling Approach 2

Forecasting Renewable Energy Generation: A Predictive Modell...

引用

2nd International Conference on Self Sustainable Artificial Intelligence Systems, ICSSAS 2024

作者： Mondal, Priotosh Bhatia, Aditi Panjwani, Roshini Dokare, Indu Ves Institute of Technology Department of Computer Engineering Mumbai India

ISBN: (纸本)9798350368413

India's rapid urbanization demands innovative solutions to address energy consumption patterns while reducing reliance on fossil fuels. This research paper explores the application of predictive modelling techniques like Multilayer Perceptron and Polynomial Regression to forecast solar energy generation and Long Short-Term Memory (LSTM) models to forecast wind energy generation, thereby facilitating efficient energy planning in major Indian cities. The proposed system aims to create a web-based platform that integrates these predictive models to display real-time temperature conditions and the corresponding amount of energy that solar and wind sources can provide in specific locations, thus promoting smart cities and smart homes. By utilizing the website, energy planners will be able to compare the generated energy from wind and solar sources, enabling informed decisions on which resource best meets the energy requirements of housing settlements. © 2024 IEEE.

关键词： Wind forecasting

来源：评论

学校读者我要写书评

暂无评论

Using Unsupervised Learning to detect Fraud in Cryptocurrency 3

Using Unsupervised Learning to detect Fraud in Cryptocurrenc...

引用

3rd IEEE International Conference on Blockchain and Distributed Systems Security, ICBDS 2024

作者： Chordia, Sankalp Nilesh Shinde, Sandeep Vishwakarma Institute of Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798350354348

Bitcoin, the most popular form of cryptocurrency, operates independently of central banks or governmental entities. The emphasis on anonymity complicates the task of tracing cryptocurrency payments linked to illicit activities or criminal transactions. Consequently, ransomware developers exploit this environment to extort ransom by encrypting valuable user information. Numerous research endeavors have explored the application of machine learning techniques to detect fraudulent activities within cryptocurrency transactions. However, there is a notable absence of studies showcasing the practical implementation of such methods using real-time blockchain data. The present study is structured into two main sections. Firstly, we delineate the development of various features utilized for training purposes. Subsequently, our approach will be validated using real-time Bitcoin data. The other section highlights the use of open source investigation tool (OSINT) to carry forward the cryptocurrency investigation by taking the input from previous findings. OSINT involves the retrieval of information from a range of publicly accessible data sources, while offering a viable approach to address the difficulties associated with extracting and aggregating intelligence from diverse publicly available information channels and social platforms. Thus narrowing down the search investigation by exploiting the pseudonymity of Bitcoin and tracing back the wallet and IP addresses' digital and physical identity. © 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

Stock Indexes Community Identification Using BAT-Modified Optimization Algorithm

引用

SN computer Science 2024年第8期5卷 1-14页

作者： Aggarwal, Kirti Arora, Anuja Department of Computer Science and Engineering and Information Technology Jaypee Institute of Information Technology Noida India

Stock price accelerates interest and preference of the young generation to explore the stock market with elicit interest. An autopilot system is needed where users choose beneficial stocks of their choice without paying attention to the manual inspection or implication. This study aims to form highly correlated stock communities based on fluctuations in stock price which will be helpful in the prediction and decision-making of users about their expenditure in the stock market. For the enormous volume of stock price time series data, the opted data representation method is the network which is generated with the help of a data statistical correlation test. Another research challenge was to detect the community in an optimized manner due to complex stock networks. For the same, BAT optimization algorithm and a proposed BAT-modified (BAT-M) algorithm are applied for complex stock network community formation. The better community formation results showcase that BAT-M gives better performance in comparison to the standard BAT algorithm on various performance evaluation factors. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： BAT Community detection Correlation Modularity Nature-inspired Optimization Stock indexes Stock network

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach for Optimal Cluster Identification and N-Order Hesitation Based Time Series Forecasting

引用

SN computer Science 2024年第7期5卷 872页

作者： Dixit, Ankit Jain, Shikha Department of Computer Science and Engineering and Information Technology Jaypee Institute of Information Technology Noida India

Time series forecasting is an important field of research, especially when the series is completely random, known as a strictly non-stationary time series (NS-TS). To handle the randomness efficiently, the paper presents a novel approach to identify the optimal number of clusters in NS-TS followed by a hesitation-based N-order IFS forecasting algorithm (called IFSH_new). The performance of the proposed approach for determining the optimal number of clusters is demonstrated using four variants of the proposed algorithm, IFSH_new, across five datasets. Furthermore, a comparison of the best variant of the proposed approach with existing state-of-art work based on mean absolute percentage error (MAPE) and root mean square error (RMSE) confirms the superiority of the proposed model over others. Additionally, the graphical representation of actual versus forecasted data points for different datasets shows good improvement at high change points compared to other approaches. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Degree of hesitation Intuitionistic fuzzy set (IFS) N-order forecasting model Non-stationary time series (NS-TS) Time series forecasting model

来源：评论

学校读者我要写书评

暂无评论

Noval Approach using Image Detection to Identify Diseases in Fruit Plants (Apple) 9

Noval Approach using Image Detection to Identify Diseases in...

引用

9th IEEE International Conference for Convergence in technology, I2CT 2024

作者： Kadam, Anup Dwivedi, Anand Prakash Shekhar, Amatya Sree, Dasari Hema Prajapat, Avnish Army Institute of Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798350394474

In the face of the critical imperative to feed a projected global population of 9 billion people by 2050, India, harboring one of the world's largest populations, grapples with formidable challenges in food production, exacerbated by substantial crop losses. An alarming statistic reveals that approximately 30 percent of India's agricultural yield succumbs to various factors, including pests, diseases, and post-harvest inefficiencies. This staggering loss not only poses a severe threat to India's food security but also contributes significantly to the broader global food crisis. Traditional solutions, often reliant on chemical pesticides, not only fall short in effectively addressing these challenges but also incur adverse environmental impacts, further exacerbating the sustainability dilemma. In response to this pressing issue, a groundbreaking research project emerges, advocating for innovative approaches that seamlessly integrate advanced technology, data-driven insights, and inclusivity to forge a resilient and sustainable agricultural ecosystem. This transformative initiative specifically targets the identification of plant diseases in apple trees, recognizing the potential ripple effects on broader agricultural practices. By leveraging cutting-edge technology like CNN and CAE the research introduces a comprehensive methodology that not only efficiently identifies and manages plant diseases but also holds the promise of sustainable crop management and economic stability. © 2024 IEEE.

关键词： Crops

来源：评论

学校读者我要写书评

暂无评论

Towards Safer Environments: Spatiotemporal Anomaly Detection and Caution System Powered by LRCN

Towards Safer Environments: Spatiotemporal Anomaly Detection...

引用

2024 International Conference on Emerging Innovations and Advanced Computing, INNOCOMP 2024

作者： Mane, Deepak Patil, Vishal Patil, Bhavesh Giri, Rakshanda Lokhande, Priyanka Vishwakarma Institute of Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798350376470

In a routine walk home from work, the urgency of personal safety was highlighted by a sudden confrontation and theft. This incident sparked the motivation for this research, with an emphasis on the imperative need for enhanced protection for all individuals. Addressing the escalating concern of crime, a system adept at detecting anomalous human activities is presented in our research, aiming to significantly bolster crime prevention measures. Integrated with a responsive Telegram bot, our solution springs into action when a detected activity exceeds a 90% confidence threshold. The nature of this activity is then evaluated by the system's decision tree: if it is classified as anomalous, the relevant authorities are immediately contacted by the bot, ensuring swift intervention. Conversely, non-anomalous activities result in non-dangerous activities. This dual-response mechanism enhances the system's versatility, making it apt for a wide range of environments, from bustling public spaces to residential areas. The proposed system's impressive accuracy rate of 92.20% in identifying malicious activity is essential to its effectiveness. This commendable precision is achieved by harnessing the UCF-Crime dataset, a comprehensive collection of surveillance footage annotated for various types of crimes. Further bolstering our system's capabilities is the integration of the Long-term Recurrent Convolutional Network (LRCN) model. LRCN stands out for its proficiency in amalgamating spatial and temporal features from video data, granting our system a heightened sensitivity in detecting subtle, yet critical, behavioural anomalies. © 2024 IEEE.

关键词： Crime

来源：评论

学校读者我要写书评

暂无评论

Twitter Disaster Prediction Using Different Deep Learning Models

引用

SN computer Science 2024年第1期5卷 179页

作者： Duraisamy, Premkumar Natarajan, Yuvaraj Department of Computer Science and Engineering KPR Institute of Engineering and Technology Tamil Nadu Coimbatore India

In emergencies, Twitter has become an essential way for people to communicate. Since many people use smartphones, they can easily report events they see. Because of this, more organizations are using Twitter to monitor what is happening digitally. Twitter and similar online platforms provide a place to share their personal experiences and local events with others. People often post about their daily lives, things happening in their neighborhoods, and their activities to keep others informed. In this study, we focus on predicting disasters on Twitter using an improved Bidirectional Encoder Depiction from the Transformers (BERT) model. Twitter has become a crucial communication tool during emergencies, allowing people to report events they witness quickly. As a result, organizations are increasingly using Twitter for digital monitoring. Our research utilizes a modified version of the BERT model, a powerful language representation technique. This modified BERT model enables us to understand better tweets related to disasters and improve the accuracy of our predictions. Using Twitter data, we train our improved BERT model to recognize patterns and signals associated with different types of disasters, such as natural calamities or accidents. We aim to efficiently detect and respond to potential disasters in real time by doing so. Our experiments' results demonstrate our BERT effectiveness, showing improved disaster prediction performance compared to traditional models like the long short-term memory model (LSTM) and gated recurrent unit model (GRU). This research contributes valuable insights into disaster management strategies and allows leveraging social media data in disaster preparedness and response efforts. © 2024, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Bidirectional encoder depiction from the transformers Gated recurrent unit model Long short-term memory model

来源：评论

学校读者我要写书评

暂无评论

Vision-Based Multilingual Sign Language Translation

引用

SN computer Science 2023年第6期4卷 715页

作者： Ghotkar, Archana Barde, Udit Sonawane, Sheetal Gokhale, Atharva Department of Computer Engineering Pune Institute of Computer Technology Maharashtra Pune 411043 India

Like spoken languages, different countries use different sign languages which include a variety of gestures and movements. This demands the need for a system that can translate one sign language to another. Thus, a computer vision-based sign-to-sign translator has been proposed which would contribute to improving the communication between the users of different sign languages. To analyze input sign language gestures, the system makes use of suitable frameworks for determining hand shapes, hand position with respect to other body parts, hand movement, and body pose key points, and can predict the Hamburg Notation System (HamNoSys) equivalent of these gestures in the form of symbols which is used as an intermediate language. Each sign language gesture is stored in HamNoSys form as a tagged dataset and the respective sign in the target sign language is played through animation by choosing the respective HamNoSys equivalent from the dataset. The proposed system can translate 66 signs in total (22 each from three sign languages) and gives an accuracy of 86.39%. The presented solution is an attempt to solve the problem of complete sign-to-sign conversion as against most previous approaches that focus on sign-to-text translation in the domain of human–computer interaction (HCI). © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： computer vision Gesture recognition HamNoSys Human–computer interaction Sign language translation

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：