检索结果-内蒙古大学图书馆

2024 IEEE International Conference on Signal Processing and Advance Research in Computing, SPARC 2024

作者： Chauhan, Shweta Kumar, Rahul Kumar, Bibek Madhav Institute of Science and Technology Department of Artificial Intelligence and Machine Learning Madhya Pradesh Gwalior474005 India GLA University ECE Department Uttar Pradesh Mathura281406 India Symbiosis University of Applied Sciences School of Data Science Madhya Pradesh Indore453112 India

ISBN: (纸本)9798350385199

Automatic age and gender classification has become important researcher since the rise of social platforms and social media. The main advancements in face recognition are the ability to determine a person's age and gender. This paper makes a significant contribution by providing a survey of different face recognition methods for determining age and gender. A deep neural network (DNN) known as a convolutional neural network is frequently used for natural language processing and image recognition and processing. The proposed method significantly enhances three elements: face alignment, face recognition, and face recognition accuracy. All weight vectors of people in the same age group within a given database are averaged together. To predict gender and age, picture classification and regression models have been discussed in this paper respectively. Although training the proposed network to make accurate predictions takes a lot of time, it significantly improves age and gender categorization precision. From the experimental results we compared and analyzed the work to the most recent UTK-Face benchmark for estimating age and gender. © 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

StaR Maps: Unveiling Uncertainty in Geospatial Relations

StaR Maps: Unveiling Uncertainty in Geospatial Relations

引用

International Conference on Intelligent Transportation

作者： Benedict Flade Simon Kohaut Julian Eggert Devendra Singh Dhami Kristian Kersting Honda Research Institute Europe GmbH Offenbach Germany Department of Computer Science Artificial Intelligence and Machine Learning Lab TU Darmstadt Darmstadt Germany Department of Mathematics and Computer Science Uncertainty in Artificial Intelligence Group TU Eindhoven Eindhoven MB Netherlands Hessian AI Centre for Cognitive Science German Center for Artificial Intelligence (DFKI)

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

The growing complexity of intelligent transportation systems and their applications in public spaces has increased the demand for expressive and versatile knowledge representation. While various mapping efforts have achieved widespread coverage, including detailed annotation of features with semantic labels, it is essential to understand their inherent uncertainties, which are commonly underrepresented by the respective geographic information systems. Hence, it is critical to develop a representation that combines a statistical, probabilistic perspective with the relational nature of geospatial data. Further, such a representation should facilitate an honest view of the data's accuracy and provide an environment for high-level reasoning to obtain novel insights from task-dependent queries. Our work addresses this gap in two ways. First, we present Statistical Relational Maps (StaR Maps) as a representation of uncertain, semantic map data. Second, we demonstrate efficient computation of StaR Maps to scale the approach to wide urban spaces. Through experiments on real-world, crowd-sourced data, we underpin the application and utility of StaR Maps in terms of representing uncertain knowledge and reasoning for complex geospatial information.

关键词： Vocabulary Uncertainty Navigation Semantics Urban areas Stars Probabilistic logic Cognition Geospatial analysis Computational efficiency

来源：评论

学校读者我要写书评

暂无评论

SARA: Structural and Adversarial Representation Alignment for Training-efficient Diffusion Models

arXiv

引用

arXiv 2025年

作者： Chen, Hesen Wang, Junyan Tan, Zhiyu Li, Hao Fudan University China Shanghai Academy of Artificial Intelligence for Science China Australian Institute for Machine Learning The University of Adelaide Australia

Modern diffusion models encounter a fundamental trade-off between training efficiency and generation quality. While existing representation alignment methods, such as REPA, accelerate convergence through patch-wise alignment, they often fail to capture structural relationships within visual representations and ensure global distribution consistency between pretrained encoders and denoising networks. To address these limitations, we introduce SARA, a hierarchical alignment framework that enforces multi-level representation constraints: (1) patch-wise alignment to preserve local semantic details, (2) autocorrelation matrix alignment to maintain structural consistency within representations, and (3) adversarial distribution alignment to mitigate global representation discrepancies. Unlike previous approaches, SARA explicitly models both intrarepresentation correlations via self-similarity matrices and inter-distribution coherence via adversarial alignment, enabling comprehensive alignment across local and global scales. Experiments on ImageNet-256 show that SARA achieves an FID of 1.36 while converging twice as fast as REPA, surpassing recent state-of-the-art image generation methods. This work establishes a systematic paradigm for optimizing diffusion training through hierarchical representation alignment. © 2025, CC BY.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Harmonizing Precision Agriculture: Augmented Insights into Plant Disease Detection Using Deep learning

Harmonizing Precision Agriculture: Augmented Insights into P...

引用

2024 International Conference on Integrated Circuits, Communication, and Computing Systems, ICIC3S 2024

作者： Chauhan, Shweta Kumar, Rahul Kumar, Bibek MITS College of Engineering Department of Artificial Intelligence and Machine Learning Gwalior474005 India GLA University ECE Department Uttar Pradesh Mathura281406 India School of Data Science Symbiosis University of Applied Sciences Madhya Pradesh Indore453112 India

ISBN: (纸本)9798350364088

This research addresses the critical issue of accurate plant disease identification, recognizing its paramount significance in preserving agricultural yields and safeguarding natural ecosystems. Leveraging the power of deep learning, specifically Convolutional Neural Networks (CNNs), our study focuses on developing a robust image classification model capable of precisely detecting various plant diseases. However, the challenge of limited training data necessitated the implementation of diverse data augmentation techniques to enrich the dataset. The research involves a meticulous exploration of different augmentation strategies, including rotation, flipping, zooming, translation, shearing, and cropping. The acquired dataset comprises 25,000 images categorized into 38 classes, with 20,000 designated for training and 5,000 for validation. The CNN model, with convolutional, pooling, dense, dropout, and flatten layers, exhibits a remarkable accuracy of 93.5% and 95% on the valid and training dataset, reflecting its adeptness in learning intricate patterns. Furthermore, the study reviews recent advancements in plant disease identification, including innovative models like PPLC-Net, Deep CNN, SG2-ADA, and HQIA. These breakthroughs underscore the continuous evolution of techniques to address challenges such as imbalanced datasets, computational demands, and overfitting. The research concludes with recommendations for future exploration, emphasizing the potential refinement of data augmentation techniques, hyperparameter tuning, and broader disease dataset acquisition. The outcomes presented here contribute to the growing body of knowledge in precision agriculture and pave the way for further advancements in plant disease detection methodologies. © 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

Hybrid Makes Better: Privacy-Preserving Medical Image Classification Based on Federated learning

Hybrid Makes Better: Privacy-Preserving Medical Image Classi...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Wang, Zhi Chi, Haotian Li, Yonggang Zhang, Yuxuan Jiang, Shunrong Du, Xiaojiang Guizani, Mohsen China University of Mining and Technology School of Computer Science Xuzhou221116 China Shanxi University School of Automation and Software Engineering Taiyuan030006 China Stevens Institute of Technology Department of Electrical and Computer Engineering HobokenNJ United States Mohamed Bin Zayed University of Artificial Intelligence Department of Machine Learning Abu Dhabi United Arab Emirates

ISBN: (纸本)9798350351255

The power of machine learning makes it available for medical image classification. However, machine learning requires large medical datasets to improve accuracy, which will involve patients' private information and lead to their privacy leakage. Federated learning is a trending technique to both protect patients' privacy and improve the accuracy of medical image classification. Unfortunately, current research shows that federated learning faces the risk of privacy leakage. In this paper, we propose a privacy-preserving scheme FHDM. Specifically, we construct a local hybrid differential privacy algorithm (LHDP) against patients' privacy leakage. This hybrid algorithm utilizes both Gaussian and Laplace differential privacy without enlarging the privacy budget. We prove that the algorithm applies to the model parameter. Moreover, we design loss optimization and global optimization strategies on the algorithm to achieve higher accuracy in medical image classification. Finally, we validate FHDM in terms of privacy-preserving and model accuracy on real datasets. Experiments show that FHDM effectively protects privacy and improves the average accuracy by 10.0% compared to adopting the LDP-based scheme with the same medical image dataset and privacy budget. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

StableMamba: Distillation-free Scaling of Large State-Space Models for Images and Videos

arXiv

引用

arXiv 2024年

作者： Suleman, Hamid Wasim, Syed Talal Naseer, Muzammal Gall, Juergen Universtiy of Bonn Germany Lamarr Institute for Machine Learning and Artificial Intelligence Germany Department of Computer Science Khalifa University Abu Dhabi United Arab Emirates

State-space models (SSMs), exemplified by S4, have introduced a novel context modeling method by integrating state-space techniques into deep learning. Despite their effectiveness, SSMs struggle with global context modeling due to data-independent matrices. The Mamba model addresses this with data-dependent variants enabled by S6 selective-scan algorithm, enhancing context modeling, especially for long sequences. However, Mamba-based architectures face significant parameter scalability challenges, limiting their utility in vision applications. This paper tackles the scalability issue of large SSMs for image classification and action recognition without relying on additional techniques like knowledge distillation. We analyze the distinct characteristics of Mamba-based and Attention-based models, proposing a Mamba-Attention interleaved architecture that enhances scalability, robustness, and performance. We demonstrate that the stable and efficient interleaved architecture resolves the scalability issue of Mamba-based architectures and increases robustness to common corruption artifacts. Our thorough evaluation on the ImageNet-1K, Kinetics-400, and Something-Something-v2 benchmarks demonstrates that our approach improves the accuracy of state-of-the-art Mamba-based architectures by up to +1.7%. © 2024, CC BY-NC-SA.

关键词： State space methods

来源：评论

学校读者我要写书评

暂无评论

Comprehensive Study and Analysis of 3D Objects from Real World Environment 6th

Comprehensive Study and Analysis of 3D Objects from Real Wor...

引用

6th International Conference on Information Systems and Management science, ISMS 2023

作者： Prajapati, Purvi Bhatt, Nirav Bhatt, Nikita Smt. K. D. Patel Department of Information Technology CSPIT Charotar University of Science and Technology Gujarat Changa India Department of Artificial Intelligence and Machine Learning CSPIT Charotar University of Science and Technology Gujarat Changa India U & P U. Patel Department of Computer Engineering CSPIT Charotar University of Science and Technology Gujarat Changa India

ISBN: (纸本)9783031707889

Over this century, deep learning has undergone some remarkable furtherance in the understanding and analysis of 2D images. Regardless of these achievements, techniques for understanding and detecting 3D sensed data, from 2D images, is relatively infantile. This literature reviews existing state-of-the-art approaches of Deep learning for detecting objects in 2D images, and detecting 3D objects from 2D images itself. We address the traditional methodologies as well as background concepts for object detection and image segmentation of 2D images eventually reviewing the latest vital approaches, comprising RGB-D, volumetric, and multi-view used for 3D objects. Datasets that are prominent for this process are also explained hereby. This paper represents a comparative analysis regarding the potential of deep learning for 3D sensed data along with its performance measures. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

GPT Assisted Annotation of Rhetorical and Linguistic Features for Interpretable Propaganda Technique Detection in News Text

arXiv

引用

arXiv 2024年

作者： Hamilton, Kyle Longo, Luca Božić, Bojan School of Computer Science and SFI Centre for Research Training in Machine Learning Technological University Dublin Dublin Ireland Artificial Intelligence and Cognitive Load Research Lab School of Computer Science and SFI Centre for Research Training in Machine Learning Technological University Dublin Dublin Ireland

While the use of machine learning for the detection of propaganda techniques in text has garnered considerable attention, most approaches focus on "black-box" solutions with opaque inner workings. Interpretable approaches provide a solution, however, they depend on careful feature engineering and costly expert annotated data. Additionally, language features specific to propagandistic text are generally the focus of rhetoricians or linguists, and there is no data set labeled with such features suitable for machine learning. This study codifies 22 rhetorical and linguistic features identified in literature related to the language of persuasion for the purpose of annotating an existing data set labeled with propaganda techniques. To help human experts annotate natural language sentences with these features, RhetAnn, a web application, was specifically designed to minimize an otherwise considerable mental effort. Finally, a small set of annotated data was used to fine-tune GPT-3.5, a generative large language model (LLM), to annotate the remaining data while optimizing for financial cost and classification accuracy. This study demonstrates how combining a small number of human annotated examples with GPT can be an effective strategy for scaling the annotation process at a fraction of the cost of traditional annotation relying solely on human experts. The results are on par with the best performing model at the time of writing, namely GPT-4, at 10x less the cost. Our contribution is a set of features, their properties, definitions, and examples in a machine-readable format, along with the code for RhetAnn and the GPT prompts and fine-tuning procedures for advancing state-of-the-art interpretable propaganda technique detection. © 2024, CC BY.

关键词： Cost engineering

来源：评论

学校读者我要写书评

暂无评论

Cross-Modal Retrieval for Motion and Text via DropTriple Loss 5

Cross-Modal Retrieval for Motion and Text via DropTriple Los...

引用

5th ACM International Conference on Multimedia in Asia, MMAsia 2023

作者： Yan, Sheng Liu, Yang Wang, Haoqiang Du, Xin Liu, Mengyuan Liu, Hong School of Artificial Intelligence Chongqing University of Technology China College of Computer Science Sichuan University China Key Laboratory of Machine Perception Shenzhen Graduate School Peking University China

ISBN: (纸本)9798400702051

Cross-modal retrieval of image-text and video-text is a prominent research area in computer vision and natural language processing. However, there has been insufficient attention given to cross-modal retrieval between human motion and text, despite its wide-ranging applicability. To address this gap, we utilize a concise yet effective dual-unimodal transformer encoder for tackling this task. Recognizing that overlapping atomic actions in different human motion sequences can lead to semantic conflicts between samples, we explore a novel triplet loss function called DropTriple Loss. This loss function discards false negative samples from the negative sample set and focuses on mining remaining genuinely hard negative samples for triplet training, thereby reducing violations they cause. We evaluate our model and approach on the HumanML3D and KIT Motion-Language datasets. On the latest HumanML3D dataset, we achieve a recall of 62.9% for motion retrieval and 71.5% for text retrieval (both based on R@10). The source code for our approach is publicly available at https://***/eanson023/rehamot. © 2023 Copyright 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

Dual-Level Contrastive learning with Reverse Translation for Summarisation

Dual-Level Contrastive Learning with Reverse Translation for...

引用

Smart Technologies for Sustainable Development Goals (ICSTSDG), International Conference on

作者： S Joanna Lois Maruboina Nithyanjali I N Nandithaa Deva Hema D Department of Computer Science Engineering with Specialization in Artificial Intelligence and Machine Learning SRM Institute of Science and Technology Ramapuram Chennai India Department of Computer Science Engineering SRM Institute of Science and Technology Ramapuram Chennai India

ISBN: (数字)9798331516284

ISBN: (纸本)9798331516291

Text summarisation aims to create brief and precise summaries from large amounts of textual information. Although recent advancements have improved summary brevity, significant room remains for enhancing ROUGE scores, especially with increase in beam width. We present a novel model, Dual Level Contrastive learning with Reverse Translation (DC-RT), which utilizes contrastive learning with data augmentation, including reverse translation, to improve accuracy. Reverse translation serves as a key augmentation strategy by translating summaries to another language and back resulting in enriched training data and enhanced robustness of the model. Additionally, we introduce a new scoring function during training to further refine output quality. Exceptional ROUGE scores are achieved using this framework, ensuring concise and readable summaries even with increased beam width. Experiments on the CNN/DailyMail (achieving a47.82 ROUGE-1 score with a 0.017 variance) dataset demonstrates that there is a significant improvement in the precision and succinctness of summaries. Notably, our model surpasses the output of BRIO, the cutting edge model, setting a new benchmark in the field.

关键词： Training Translation Training data Text summarization Contrastive learning Benchmark testing Data augmentation Data models Robustness Sustainable development

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：