检索结果-内蒙古大学图书馆

Ultrasound-based Kidney Disease Detection: Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Network 8

学校读者我要写书评

暂无评论

Ultrasound-based Kidney Disease Detection: Resampling Cubatu...

8th International Conference on Inventive Computation Technologies, ICICT 2025

作者： Padmanayana Kurumeti, N.S.L. Kumar Veena, S. Raja, Ch. Upadhye, Vijay Jagdish Maranan, Ramya Department of Artificial Intelligence and Data Science Engineering AJ Institute of Engineering and Technology Karnataka Mangaluru 575006 India Department of Artificial Intelligence and Machine Learning Aditya University Andhra Pradesh Surampalem 533437 India Department of Electronics and Communication Engineering SJC Institute of Technology Karnataka Chikkaballapur 562101 India Department of Electronics and Communication Engineering Mahatma Gandhi Institute of Technology Telangana Hyderabad 500075 India Department of Microbiology Research and Development Cell (RDC) Dept of Microbiology Parul Institute of Applied Sciences (PIAS) Parul University Post Limda Gujarat Waghodia 391760 India Department of Research and Innovation Saveetha School of Engineering SIMATS Tamil Nadu Chennai 602105 India

ISBN: (纸本)9798331512248

Limitations of ultrasound-based identification of kidney disease include poor image resolution, operator reliance, and lack of detection of small tissue anomalies. Diagnostic accuracy may be affected by artefacts or noise, or changes in patient anatomy. It may also have difficulty detecting microscopic structural changes or renal disease in its incipiency. The limitations are overcome by the analysis of the Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Networks (RCKF-PG-VGG-CCNet) for kidney disease detection from ultrasound images based on the Kidney Ultrasound Image Dataset. The initial use of the Resampling Cubature Kalman Filter and the first use of Physics-Guided VGG Cross-Contextual Networks is to enhance image quality. Through identifying regions through deviations, the Scale-Aware Meet Transformer (SAMT) performs segmentation. While the Synergistic Fibroblast Optimization (SFO) enhances classification performance through optimization of the search for diseased regions, the Physics-Guided VGG Cross-Contextual Networks (PG-VGG-CCNet) classify kidney disease and identify features. Experimental output from a Python-based experiment demonstrates that RCKF-PG-VGG-CCNet is more efficient than other methods with 99.9% efficiency and 99.8% sensitivity and may enhance the accuracy of kidney disease detection. © 2025 IEEE.

关键词： Kidney disease Kidney Ultrasound Image Dataset Resampling Cubature Kalman Filter Scale-Aware Meet Transformer Synergistic Fibroblast Optimization

artificial Electromagnetic Medium Usages in the Energy Sector for the Purposes of Transmission

学校读者我要写书评

暂无评论

Artificial Electromagnetic Medium Usages in the Energy Secto...

International Conference on Power Energy, Environment and Intelligent Control (PEEIC)

作者： Manju Bargavi Subhash Chandra Nishant Saxena Department of Computer Jain (Deemed to be University) Bangalore India Department of Artificial Intelligence & Machine Learning Noida Institute Of Engineering and Technology Greater Noida Uttar Pradesh India Department of Electronics and Communication Engineering Tula's Institute Dehradun India

We provide a thorough investigation of the field of metamaterials in order and meta surfaces with particular attention to its uses in wireless energy extraction and wireless power transfer (WPT). In WPT, the electromagnetic field path from source to receiver is divided by radiative technologies, which include both near- and far-field methods, and nonradiative near-field technologies. We evaluate a wide range of designs, comparing and contrasting their many aspects. Our research shows that meta material and meta surfaces are essential for increasing the effectiveness and operating diversity of WPT systems. These novel materials may reduce sensitivity to incoming wave angles and polarizations, increasing the energy conversion effectiveness in wireless energy harvesters. A vital component of both WPT und WEH is the rectenna. Our results demonstrate the advantages of rectennas based on metamaterials, demonstrating their capacity to attain better conversion efficiencies from RF to DC. These prospects highlight the dynamic character of this area, ranging from researching innovative applications to enhancing current solutions. With the advent of wireless energy collecting and transmission, metamaterials play an increasingly important role in influencing the development of efficient and sustainable technologies.

关键词：

Streamlining Attendance with Fast Face Recognition Model

学校读者我要写书评

暂无评论

Streamlining Attendance with Fast Face Recognition Model

International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)

作者： Koushik Sundar Ravikumar S Jeyalakshmi J Eugene Berna I Prithi Samuel Citibank NA Jersey City NJ USA Department of Computer Science & Engineering Vel Tech Rangarajan Dr.Sagunthala R&D Institute of Science and Technology Chennai Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidhyapeetham Chennai India Department of Artificial Intelligence and Machine Learning Bannari Amman Institute of Technology Erode India Department of Computational Intelligence School of Computing SRM Institute of Science and Technology Chennai India

ISBN: (数字)9798350376425

ISBN: (纸本)9798350376432

Facial recognition technology has become increasingly prevalent in today's world, with applications ranging from mobile phones to banking and workplace security. This research proposes a facial recognition-based solution to replace the manual attendance tracking process, streamlining the process and reducing human error. The proposed system integrates facial recognition, registration forms, and a Redis database to automate attendance tracking. By leveraging the ML search algorithm (similarity approach) and the Insight face Python module, the system can accurately identify individuals within a database. The system's implementation involves two phases: the initial phase focuses on integrating facial recognition, registration forms, and database access. The second phase involves utilizing software such as Streamlit and Dashboard to enhance user experience and visualization. This innovative solution offers a more efficient and accurate alternative to traditional manual attendance tracking methods, improving overall efficiency and reducing administrative burdens. By automating the process, the system contributes to a more streamlined and secure environment.

关键词： Training Accuracy Databases Face recognition Manuals Feature extraction User experience Security Usability System implementation

Preventing Representational Rank Collapse in MPNNs by Splitting the Computational Graph

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Roth, Andreas Bause, Franka Kriege, Nils M. Liebig, Thomas Faculty of Computer Science TU Dortmund University Dortmund Germany Faculty of Computer Science University of Vienna Vienna Austria UniVie Doctoral School Computer Science University of Vienna Vienna Austria Research Network Data Science University of Vienna Vienna Austria Lamarr Institute for Machine Learning and Artificial Intelligence Dortmund Germany

The ability of message-passing neural networks (MPNNs) to fit complex functions over graphs is limited as most graph convolutions amplify the same signal across all feature channels, a phenomenon known as rank collapse, and oversmoothing as a special case. Most approaches to mitigate over-smoothing extend common message-passing schemes, e.g., the graph convolutional network, by utilizing residual connections, gating mechanisms, normalization, or regularization techniques. Our work contrarily proposes to directly tackle the cause of this issue by modifying the message-passing scheme and exchanging different types of messages using multi-relational graphs. We identify a sufficient condition to ensure linearly independent node representations. As one instantion, we show that operating on multiple directed acyclic graphs always satisfies our condition and propose to obtain these by defining a strict partial ordering of the nodes. We conduct comprehensive experiments that confirm the benefits of operating on multi-relational graphs to achieve more informative node representations. Copyright © 2024, The Authors. All rights reserved.

关键词： Convolution

STRAP: ROBOT SUB-TRAJECTORY RETRIEVAL FOR AUGMENTED POLICY learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Memmel, Marius Berg, Jacob Chen, Bingqing Gupta, Abhishek Francis, Jonathan Paul G. Allen School of Computer Science & Engineering University of Washington United States Robot Learning Lab Bosch Center for Artificial Intelligence United States Robotics Institute Carnegie Mellon University United States

Robot learning is witnessing a significant increase in the size, diversity, and complexity of pre-collected datasets, mirroring trends in domains such as natural language processing and computer vision. Many robot learning methods treat such datasets as multi-task expert data and learn a multi-task, generalist policy by training broadly across them. Notably, while these generalist policies can improve the average performance across many tasks, the performance of generalist policies on any one task is often suboptimal due to negative transfer between partitions of the data, compared to task-specific specialist policies. In this work, we argue for the paradigm of training policies during deployment given the scenarios they encounter: rather than deploying pre-trained policies to unseen problems in a zero-shot manner, we non-parametrically retrieve and train models directly on relevant data at test time. Furthermore, we show that many robotics tasks share considerable amounts of low-level behaviors and that retrieval at the "sub"-trajectory granularity enables significantly improved data utilization, generalization, and robustness in adapting policies to novel problems. In contrast, existing full-trajectory retrieval methods tend to underutilize the data and miss out on shared cross-task content. This work proposes STRAP, a technique for leveraging pre-trained vision foundation models and dynamic time warping to retrieve sub-sequences of trajectories from large training corpora in a robust fashion. STRAP outperforms both prior retrieval algorithms and multi-task learning methods in simulated and real experiments, showing the ability to scale to much larger offline datasets in the real world as well as the ability to learn robust control policies with just a handful of real-world demonstrations. Project website at https://***/strap/. © 2024, CC BY.

关键词： Multi-task learning

PROTLLM: An Interleaved Protein-Language LLM with Protein-as-Word Pre-Training

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhuo, Le Chi, Zewen Xu, Minghao Huang, Heyan Zheng, Heqi He, Conghui Mao, Xian-Ling Zhang, Wentao School of Computer Science and Technology Beijing Institute of Technology China Beihang University China Center for Machine Learning Research Peking University China State Grid Smart Grid Research Institute Co. Ltd China Shanghai Artificial Intelligence Laboratory China

We propose PROTLLM, a versatile cross-modal large language model (LLM) for both protein-centric and protein-language tasks. PROTLLM features a unique dynamic protein mounting mechanism, enabling it to handle complex inputs where the natural language text is interspersed with an arbitrary number of proteins. Besides, we propose the protein-as-word language modeling approach to train PROTLLM. By developing a specialized protein vocabulary, we equip the model with the capability to predict not just natural language but also proteins from a vast pool of candidates. Additionally, we construct a large-scale interleaved protein-text dataset, named InterPT, for pre-training. This dataset comprehensively encompasses both (1) structured data sources like protein annotations and (2) unstructured data sources like biological research papers, thereby endowing PROTLLM with crucial knowledge for understanding proteins. We evaluate PROTLLM on classic supervised protein-centric tasks and explore its novel protein-language applications. Experimental results demonstrate that PROTLLM not only achieves superior performance against protein-specialized baselines on protein-centric tasks but also induces zero-shot and in-context learning capabilities on protein-language tasks. Copyright © 2024, The Authors. All rights reserved.

关键词： Proteins

A Review of machine learning and Deep learning algorithms for Chronic Kidney Disease Prediction

学校读者我要写书评

暂无评论

A Review of Machine Learning and Deep Learning algorithms fo...

artificial intelligence in Education and Industry 4.0 (IDICAIEI), DMIHER International Conference on

作者： Pratik Nagre Prateek Verma Praveen Kumar Nitin L. Gavankar Department of Artificial Intelligence and Machine Learning Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research Sawangi Wardha Maharashtra India Department of Computer Science and Medical Engineering Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research Sawangi Wardha Maharashtra India Department of Computer Science and Engineering Walchand College of Engineering Sangli Maharashtra India

ISBN: (数字)9798331528713

ISBN: (纸本)9798331528720

Chronic Kidney Disease (CKD) is a common disease worldwide that is manifested by slow progressive loss of kidney function that can result in complications and kidney failure. That is why conventional approaches to CKD diagnosis and treatment have certain difficulties: late identification of the disease, inaccuracy of the diagnosis, and high costs for the therapy. The new approaches for solving these problems are opened with the help of machine learning (ML) and deep learning (DL). These technologies can help existing information sets in large volumes in diagnosing, predicting, and in early detection of diseases to enhance the well-being of the patients. However, the outcome comes with a handful of constraints, including the quality of the data used in developing the models, the ability to interpret results from the generated model, and the complexity of the computation process. The following section describes the various ML and DL algorithms used for CKD, their strengths, weaknesses, and the possibility of implementing these technologies in clinical settings for better CKD outcomes.

关键词： Deep learning machine learning algorithms Accuracy Computational modeling Medical treatment learning (artificial intelligence) Prediction algorithms Chronic kidney disease Data models Diseases

Text to Image Generation Using Attentional Generative Adversarial Network 1

学校读者我要写书评

暂无评论

1st International Conference on Advanced Computing, machine learning, Robotics and Internet Technologies, AMRIT 2023

作者： Rao, Supriya B. Shetty, Shailesh S. Singh, Chandra M, Srinivas P. Bekal, Anush Department of Artificial Intelligence and Machine Learning New Horizon College of Engineering Bangalore Karnataka Mangaluru India Department of Computer Science and Business System Srinivas Institute of Technology Karnataka Mangaluru India Department of Electronics & Communication Engineering Sahyadri College of Engineering & Management Karnataka Mangaluru India Department of Computer Science Engineering Sahyadri College of Engineering & Management Karnataka Mangaluru India

ISBN: (数字)9783031472213

ISBN: (纸本)9783031472206

AttnGan – Attentional Generative Adversarial Network (GAN) is a type of GANs that is used for text to image generation. Unlike other art models, it focuses on fine-grained word level information and once the model is trained by mapping the images and the corresponding text in the dataset, the model will generate image part-by-part using the image description given by the user as an input to the model. Current paper, focuses on the conversion of text into images based on the description provided by the user. This enables them to translate their ideas in the form of text into images of their required choices. The model is initially trained by the description of the text and images and when the user provides the text as an input to the system, it is used as a testing data. By analyzing the text word-by-word, the respective images pertaining to the text is generated part-by-part by considering the text description into account. This can heavily help the users in a applications such as in editing aids, documental archives, etc. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Generative adversarial networks

Bandit Algorithms for Prophet Inequality and Pandora’s Box

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gatmiry, Khashayar Kesselheim, Thomas Singla, Sahil Wang, Yifan Electrical Engineering and Computer Science Massachusetts Institute of Technology United States Institute of Computer Science Lamarr Institute for Machine Learning and Artificial Intelligence University of Bonn Germany School of Computer Science Georgia Tech United States

The Prophet Inequality and Pandora’s Box problems are fundamental stochastic problem with applications in Mechanism Design, Online Algorithms, Stochastic Optimization, Optimal Stopping, and Operations Research. A usual assumption in these works is that the probability distributions of the n underlying random variables are given as input to the algorithm. Since in practice these distributions need to be learned under limited feedback, we initiate the study of such stochastic problems in the Multi-Armed Bandits model. In the Multi-Armed Bandits model we interact with n unknown distributions over T rounds: in round t we play a policy x(t) and only receive the value of x(t) as feedback. The goal is to minimize the regret, which is the difference over T rounds in the total value of the optimal algorithm that knows the distributions vs. the total value of our algorithm that learns the distributions from the limited feedback. Our main results give near-optimal Oe(poly(n)√T ) total regret algorithms for both Prophet Inequality and Pandora’s Box. Our proofs proceed by maintaining confidence intervals on the unknown indices of the optimal policy. The exploration-exploitation tradeoff prevents us from directly refining these confidence intervals, so the main technique is to design a regret upper bound function that is learnable while playing low-regret Bandit policies. Copyright © 2022, The Authors. All rights reserved.

关键词： Operations research