检索结果-内蒙古大学图书馆

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Chenyang Zhang Difan Zou Yuan Cao Department of Statistics and Actuarial Science School of Computing and Data Science The University of Hong Kong Department of Computer Science School of Computing and Data Science & Institute of Data Science The University of Hong Kong Department of Statistics and Actuarial Science School of Computing and Data Science & Department of Mathematics The University of Hong Kong

ISBN: (纸本)9798331314385

Adam has become one of the most favored optimizers in deep learning problems. Despite its success in practice, numerous mysteries persist regarding its theoretical understanding. In this paper, we study the implicit bias of Adam in linear logistic regression. Specifically, we show that when the training data are linearly separable, the iterates of Adam converge towards a linear classifier that achieves the maximum ℓ∞ -margin in direction. Notably, for a general class of diminishing learning rates, this convergence occurs within polynomial time. Our result shed light on the difference between Adam and (stochastic) gradient descent from a theoretical perspective.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Alignment Work for Urban Accessibility: A Study of How Wheelchair Users Travel in Urban Spaces

引用

Proceedings of the ACM on Human-Computer Interaction 2022年第CSCW2期6卷 1–22页

作者： Wu, Yiying Ding, Xianghua Dai, Xuelan Zhang, Peng Lu, Tun Gu, Ning School of Computer Science Shanghai Key Laboratory of Data Science Fudan University Shanghai China School of Computing Science University of Glasgow Glasgow United Kingdom

Mobility is the key for people with disabilities to have full participation in life. To support their mobility, previous work primarily focused on accessibility as an attribute of the external environment to be evaluated, labeled, visualized, and improved. Relatively less work has looked at how people with disabilities go out and move around in practice and work to achieve mobility. This paper presents a qualitative study of 14 wheelchair users' travel practices in everyday life in China, highlighting urban accessibility as alignment work-the use of an awareness of the conditions along the journey to continuously align and re-Align various resources to ensure accessibility and mobility across contexts and moments. By highlighting alignment work, we present another side of the story of urban accessibility, and broaden design considerations to support mobility for wheelchair users. © 2022 ACM.

关键词： Alignment

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Exploration of Real-Time 3-D View Reconstruction Methods

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第12期5卷 5915-5927页

作者： Agrawal, Arya Sharma, Teena Verma, Nishchal K. Vellore Institute of Technology School of Computer Science and Engineering Vellore632014 India Indian Institute of Technology Mehta Family School of Data Science and Artificial Intelligence Guwahati781039 India Indian Institute of Technology Kanpur Department of Electrical Engineering 208016 India

Real-time 3-D view reconstruction in an unfamiliar environment poses complexity for various applications due to varying conditions such as occlusion, latency, precision, etc. This article thoroughly examines and tests contemporary methodologies addressing challenges in 3-D view reconstruction. The methods being explored in this article are categorized into volumetric and mesh, generative adversarial network based, and open source library based methods. The exploration of these methods undergoes detailed discussions, encompassing methods, advantages, limitations, and empirical results. The real-time testing of each method is done on benchmarked datasets, including ShapeNet, Pascal 3D+, Pix3D, etc. The narrative highlights the crucial role of 3-D view reconstruction in domains such as robotics, virtual and augmented reality, medical imaging, cultural heritage preservation, etc. The article also anticipates future scopes by exploring generative models, unsupervised learning, and advanced sensor fusion to increase the robustness of the algorithms. © 2020 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Colour Correction and Detail Enhancement in Underwater Images using Hybrid Real-ESRGAN 3

Colour Correction and Detail Enhancement in Underwater Image...

引用

3rd International Conference on Innovative Mechanisms for Industry Applications, ICIMIA 2023

作者： Chandar, Vijay Nair, Vishnu G Rajeswari, D. SRM Institute of Science and Technology Kattankulathur School of Computing College of Engineering and Technology Department of Data Science and Business Systems Chennai603203 India

ISBN: (纸本)9798350343632

The motive of underwater image enhancement is to improve and enhance the quality of photographs taken underwater. Various traditonal image processing techniques and deep learning models have been developed to work on this task. But most often the deep learning models and conventional methods of image processing work in silos, leading to subpar results. This research proposes a strategy that utilizes a combination of both approaches to solve this issue. Initially, traditional image processing techniques like colour correction and image stretching are used to enhance the colour of the images. The enhanced images are then fed into a custom Real-ESRGAN model that has been fine-tuned on an underwater image dataset to optimize performance in this context. This hybrid approach has the potential to outperform the models utilizing either traditional or deep learning methods in isolation by leveraging the strengths of both. In addition, batch processing is used in the pipeline to efficiently handle multiple images at the same time. This effectively addresses the scalability problem that arises frequently in image enhancement tasks. With eye-catching visuals, the proposed methodology achieves a Peak Signal-to-Noise Ratio (PSNR) of 23.66 and a Mean Squared Error (MSE) of 0.28, suggesting excellent image restoration and noise reduction. The qualitative assessments show that the improved image graphs have better colour, contrast, and clarity than the originals. © 2023 IEEE.

关键词： Batch Processing Image Enhancement Image Stretching MSI PSNR Real-ESRGAN

来源：评论

学校读者我要写书评

暂无评论

MedEx: Enhancing Medical Question-Answering with First-Order Logic based Reasoning and Knowledge Injection 31

MedEx: Enhancing Medical Question-Answering with First-Order...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Zafar, Aizan Mishra, Kshitij Ekbal, Asif Department of Computer Science and Engineering Indian Institute of Technology Patna India School of AI and Data Science Indian Institute of Technology Jodhpur India

ISBN: (纸本)9798891761964

In medical question-answering, traditional knowledge triples often fail due to superfluous data and their inability to capture complex relationships between symptoms and treatments across diseases. This limits models' ability to provide accurate, contextually relevant responses. To overcome this, we introduce MedEx, which employs First-Order Logic (FOL)-based reasoning to model intricate relationships between diseases and treatments. We construct FOL-based triplets that encode the interplay of symptoms, diseases, and treatments, capturing not only surface-level data but also the logical constraints of the medical domain. MedEx encodes the discourse (questions and context) using a transformer-based unit, enhancing context comprehension. These encodings are processed by a Knowledge Injection Cell that integrates knowledge graph triples via a Graph Attention Network. The Logic Fusion Cell then combines medical-specific logical rule triples (e.g., co-occurrence, causation, diagnosis) with knowledge triples and extracts answers through a feed-forward layer. Our analysis demonstrates MedEx's effectiveness and generalization across medical question-answering tasks. By merging logical reasoning with knowledge, MedEx provides precise medical answers and adapts its logical rules based on training data nuances. © 2025 Association for Computational Linguistics.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

Cursive Hand writing Recognition Using Neural Networks with VGG-16 2

Cursive Hand writing Recognition Using Neural Networks with ...

引用

2nd IEEE International Conference on Advances in computing, Communication and Applied Informatics, ACCAI 2023

作者： Rangari, Anushka Anil Das, Swagat Rajeswari, D. College of Engineering and Technology Srm Institute of Science and Technology School of Computing Department of Data Science and Business Systems Tamilnadu Kattankulathur603203 India

ISBN: (纸本)9798350315905

Recognition of cursive handwritten text from photographs is a method for finding handwritten text. Identification is difficult because every author has a distinctive writing style. The proposed work uses cutting-edge technologies including RNN, KNN, MLP, and CRNN twith VGG16 to push the boundaries of character and handwriting recognition. Additionally, it compares neural networks to CNNwith VGG16 and finds that neural networks are more accurate. In the first stage, image acquisition, the scanned image is acquired along with its normalization, feature extraction, and segmentation. Additionally, it combines with the most effective feature extraction technique to improve the accuracy of our model. It also compares the accuracy with six uniquepretreat. approach, which are image segmentation image data generator data augmentation closed loop pre-processing, sigmoid Stretching and geometric *** provides the highest accuracy (98.36%) when using data augmentation. CNN has the highest accuracy (95.25%) when using picture segmentation. The third experiment's results for the image data generator show that CNN produced the best results (92.47%), while the highest model, RNN, gave closed loop processing accuracy of 92.20%. a change in geometry KNN has the highest accuracy at 95.39%. © 2023 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

An Interpretable Deep Learning-based Model for Decision-making through Piecewise Linear Approximation

引用

ACM Transactions on Knowledge Discovery from data 2025年第3期19卷 1-35页

作者： Guo, Mengzhuo Zhang, Qingpeng Zeng, Daniel Dajun Business School Sichuan University Chengdu China Musketeers Foundation Institute of Data Science The University of Hong Kong Hong Kong Institute of Automation Chinese Academy of Sciences Beijing China

Full-complexity machine learning models, such as the deep neural network, are non-traceable black-box, whereas the classic interpretable models, such as linear regression models, are often over-simplified, leading to lower accuracy. Model interpretability limits the application of machine learning models in management problems, which requires high prediction performance, as well as the understanding of individual features' contributions to the model outcome. To enhance model interpretability while preserving good prediction performance, we propose a hybrid interpretable model that combines a piecewise linear component and a nonlinear component. The first component describes the explicit feature contributions by piecewise linear approximation to increase the expressiveness of the model. The other component uses a multi-layer perceptron to increase the prediction performance by capturing the high-order interactions between features and their complex nonlinear transformations. The interpretability is obtained once the model is learned in the form of shape functions for the main effects. We also provide a variant to explore the higher-order interactions among features. Experiments are conducted on synthetic and real-world datasets to demonstrate that the proposed models can achieve good interpretability by explicitly describing the main effects and the interaction effects of the features while maintaining state-of-the-art accuracy. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Battery Health Prediction Using Deep Hybrid Learning 3

Battery Health Prediction Using Deep Hybrid Learning

引用

3rd International Conference on Pervasive computing and Social Networking, ICPCSN 2023

作者： Nair, Rahul Anil Suresh, Dev Karan Rajeswari, D. School of Computing College of Engineering and Technology Srm Institute of Science and Technology Department of Data Science and Business Systems Tamil Nadu Kattankulathur India

ISBN: (纸本)9798350322842

The type and condition of the road the vehicle travels;how long the battery is used for are just a few of the variables that affect the battery life of electric cars (EVs). These elements might affect the battery's lifespan by causing it to drain more quickly than usual. By understanding the various factors that impact battery life, work can be done to mitigate these issues and extend the useful life of the battery. To find the best circumstances for increasing the battery's longevity, this research paper is concentrated on examining the impacts of road conditions on battery life. The previous article classified routes with an accuracy of 83.56% based on battery health while arriving at the destination [1]. This article examines EV routed and categorises them on the basis of the number of brakes applied upon arrival at the destination. By doing so, it aims to develop recommendations for drivers on how to optimize their driving habits to extend the battery's lifespan. Observations indicate that bad routes generally involve a lot more brakes as compared to good routes. Vehicle speed has a negative correlation with heater power in a bad route and a positive correlation in a good route. Battery voltage and Battery state- of-charge have opposing correlations in both categories. In this research, a prediction of the route category was attempted by using Deep Hybrid Learning to integrate three pre-trained models., VGGI6, Resnet50 and VGG19 with two boosting algorithms:- XGBoost and AdaBoost. The accuracy of the six deep hybrid learning models in predicting the category of these routes are as follows:-0.87625, 0.84875, 0.80125, 0.7925, 0.8425 and 0.855. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Music Recommendation based on Facial Expressions using data Augmentation 3

Music Recommendation based on Facial Expressions using Data ...

引用

3rd International Conference on Innovative Mechanisms for Industry Applications, ICIMIA 2023

作者： Kapoor, Tanisha Ganguly, Arnaja Rajeswari, D. SRM Institute of Science and Technology Kattankulathur School of Computing College of Engineering and Technology Department of Data Science and Business Systems Chennai603203 India

ISBN: (纸本)9798350343632

Emotion recognition from facial expressions plays a pivotal role in human-computer interaction, psychology, marketing, and various other domains. This research paper presents a comprehensive study and implementation of an automated emotion recognition system based on deep learning techniques. The project addresses the challenge of accurately classifying a wide range of emotions from facial images in real-time. Leveraging Convolutional Neural Networks (CNNs) and modern deep learning approaches, the system demonstrates high accuracy in identifying emotions, including happiness, sadness, anger, and more. This project used the image dataset FER-2013 for the mood prediction portion of the project, and Spotify music data to predict emotions for music recommendation. Both the datasets utilized are freely available on Kaggle. The proposed system can achieve an accuracy of 91.23% in predicting the users' emotion based on their facial expressions. © 2023 IEEE.

关键词： CNN Emotion recognition Facial expression analysis FER-2013 Music Recommendation

来源：评论

学校读者我要写书评

暂无评论

Predicting the Impact of Road Conditions on Battery Health Via Transfer Learning 3

Predicting the Impact of Road Conditions on Battery Health V...

引用

3rd International Conference on Artificial Intelligence and Smart Energy, ICAIS 2023

作者： Nair, Rahul Anil Suresh, Dev Karan Rajeswari, D. Srm Institute of Science and Technology School of Computing College of Engineering and Technology Department of Data Science and Business Systems Tamil Nadu Kattankulathur India

ISBN: (纸本)9781665462167

Battery life of EV (electric vehicles) depends on various factors. These factors include temperature of the battery, duration of use or the type and condition of the road traversed by the vehicle. This may cause the battery to discharge faster than it used to. Understanding what may cause the battery life to shorten can help us alleviate the given problem and can preserve the battery for longer time. This research focuses on the effects road conditions have on battery life to be able to recommend the favourable conditions to extend it. It proposes to find the routes taken by the vehicles and classifying them based on their battery health upon reaching the destination. In this paper, an attempt has been made to predict the category of the routes by employing Transfer learning via a pre-trained model VGG-16. Observations indicate that paths which have more turns and curves are generally worse for battery life as compared to straight and narrow paths. Another model which is an artificial neural network has been developed to predict the based on the average vehicle speed and the distance covered by the vehicle. The VGG-16 model predicts the data with which it is validated with an accuracy of 84.29% and the testing data with an accuracy of 83.56%. The artificial neural network model predicts the category of the data with an accuracy of 88.02%. © 2023 IEEE.

关键词： Deep 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：