检索结果-内蒙古大学图书馆

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Zhang, Mingkun Li, Jianing Chen, Wei Guo, Jiafeng Cheng, Xueqi CAS Key Laboratory of AI Safety Institute of Computing Technology Chinese Academy of Sciences Beijing China Key Laboratory of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9781643685489

Adversarial purification is one of the promising approaches to defend neural networks against adversarial attacks. Recently, methods utilizing diffusion probabilistic models have achieved great success for adversarial purification in image classification tasks. However, such methods fall into the dilemma of balancing the needs for noise removal and information preservation. This paper points out that existing adversarial purification methods based on diffusion models gradually lose sample information during the core denoising process, causing occasional label shift in subsequent classification tasks. As a remedy, we suggest to suppress such information loss by introducing guidance from the classifier confidence. Specifically, we propose Classifier-cOnfidence gUided Purification (COUP) algorithm, which purifies adversarial examples while keeping away from the classifier decision boundary. Experimental results show that COUP can achieve better adversarial robustness under strong attack methods. © 2024 The Authors.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Fire Detection System using Deep Learning Techniques 4

Fire Detection System using Deep Learning Techniques

引用

4th International Conference on Inventive Research in computing Applications, ICIRCA 2022

作者： Srividhya, S. Sindhu, S. Sornalakshmi, K. Hemavathi, D. Rajaram, V. School of Computing Srm Institute of Science and Technology Data Science and Business Systems Chennai India School of Computing Srm Institute of Science and Technology Department of Networking and Communications Chennai India

ISBN: (数字)9781665497077

ISBN: (纸本)9781665497077

An image-based fire detection system is designed using an image dataset to train the model, and further, any image capturing device will be used for later prediction purposes. The training and prediction are done using deep learning techniques. Convolution Neural Network is used for implementation. As another fire identification innovation, picture fire location has as of late assumed a pivotal part in diminishing property misfortune and harm during a fire by disturbing clients at a beginning phase through early fire discovery. Picture fire location depends on an algorithmic investigation of pictures. Nonetheless, there is a lower precision, postponed recognition, and much calculation in like manner identification calculations, including physically and machine naturally separating picture highlights. The model is designed to alert the user via SMS and email along with a captured image of fire when a fire occurs. The proposed system in development aims at providing a detection system that can alert a user instantly at the time of a hazardous fire. The aim is to create an accurate model that does not result in false alerts. The model is trained using accurate image datasets to differentiate between a fire and a non-fire scenario. Existing models used the RGB-HIS colour space for their model. Particularly, it is used to obtain pre-detected areas rather than the traditional motion differential method. The centroid variety of candidates is used to measure the flame colour dispersions in subsequent frames. Multi-expert systems have the following;dispersion of colour and centroid movement are entrenched to classify fire based on testing various clips acquired by a surveillance camera in private confinement. © 2022 IEEE.

关键词： Fires

来源：评论

学校读者我要写书评

暂无评论

Winning with Losses: The Surprising Success of Negative Strategies in Social Interaction Behavior

引用

Physical Review Letters 2024年第16期133卷 167401页

作者： Joel Weijia Lai Kang Hao Cheong Institute for Pedagogical Innovation Research and Excellence School of Physical and Mathematical Sciences College of Computing and Data Science (CCDS)

In the vast network of social interactions and behaviors, it is common to find certain schemes that, on their own, might appear counterproductive to societal progress. When observed in isolation, these schemes often seem to hinder more than they help. However, due to society’s complexity, the hidden potential of combining these seemingly detrimental schemes often goes unnoticed. Here, we investigate two such social behaviors, reputation and reciprocity, and their role in explaining Darwin’s survival of the fittest, examining how these fundamental principles govern individual interactions and shape broader social dynamics. We outline the dynamics of these two social behaviors and underline the importance of combined strategies in enhancing group welfare and contributing to interdisciplinary research in social physics.

关键词： Animal behavior, social interactions & emergent phenomena Ecological networks Chaos & nonlinear dynamics

来源：评论

学校读者我要写书评

暂无评论

Clear Text Visibility: A CNN-LSTM Approach for Frame of Interest Identification in Book Flipping Videos

Clear Text Visibility: A CNN-LSTM Approach for Frame of Inte...

引用

2024 IEEE International Conference on Modeling, Simulation and Intelligent computing, MoSICom 2024

作者： Buddhawar, Gaurav Dave, Dharin Jariwala, Krupa Chattopadhyay, Chiranjoy Sardar Vallabhbhai National Institute of Technology Dept. of Computer Science and Engineering Surat India Pune India Flame University School of Computing and Data Sciences Pune India

ISBN: (纸本)9798331533311

Book flipping videos present a distinctive challenge for information extraction, requiring the identification of frames with clear text visibility during dynamic page turns. This paper introduces a novel approach to frame classification in book flipping videos, leveraging the combined effect of Convolutional Neural Networks (CNNs) and Long Short-Term Memory networks (LSTMs). In particular, a CNN-LSTM model is employed, where a pre-trained CNN captures spatial features, and LSTMs extend the model's capability to distinguish temporal dependencies critical for page turn detection. Unlike conventional methods, this approach processes entire video sequences, enabling the model to learn intricate Frame of Interest (FoI) detection features indicative of page turns and text visibility. The proposed model offers improved interpretability, effectively classifying frames without compromising transparency. The research opens avenues for automated book digitization, digital library creation, and educational technology applications, enhancing information extraction and reliability from book flipping videos. © 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

Exploring the Efficacy and Limitations of Histogram-Based Fake Image Detection 28th

Exploring the Efficacy and Limitations of Histogram-Based Fa...

引用

28th International Conference on Knowledge Based and Intelligent information and Engineering Systems, KES 2024

作者： Hölscher, Dirk Reich, Christoph Gut, Frank Knahl, Martin Clarke, Nathan Institute for Data Science Cloud Computing and IT Security Furtwangen University Furtwangen78120 Germany Centre for Cyber Security Communications and Network Research Plymouth University Plymouth United Kingdom

Generative image models pose challenges to image authenticity and trustworthiness, blurring the line between real and fake content. This paper addresses these concerns by proposing a histogram-based approach using pre-trained models (vgg16, ResNet50, Xception) to train classification networks for distinguishing real from generated images. Leveraging histograms derived from images, the method aims to accurately classify images as authentic or synthetic. Through experiments, the paper examines the effectiveness of the approach in mitigating the risks associated with fake contents widespread dissemination. Results demonstrate promising advancements in detecting image manipulation and preserving the integrity of visual information amidst the spread of generative models. Using pre-trained models the paper shows high classification accuracy for detecting fake images. © 2024 The Authors.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Stock Market Prediction Using Machine Learning Based Techniques

Stock Market Prediction Using Machine Learning Based Techniq...

引用

2024 IEEE International Conference on Intelligent Signal Processing and Effective Communication Technologies, INSPECT 2024

作者： Dubey, Aditya Sharma, Ramnaresh Bisen, Dhananjay Venu, Nookala Jaiswal, Varshali Raghuwanshi, Bhagat Singh Shakywar, Khemchand Madhav Institute of Technology & Science Centre for Internet of Things Gwalior India Madhav Institute of Technology & Science Centre for AI Gwalior India ABV Indian Institute of Information Technology and Management Data Science and Quantum Computing Gwalior India School of Engineering Avantika University MP Ujjain India

ISBN: (纸本)9798350379525

Many methods have been used to forecast stock market trends in the big data age, including real numbers, fuzzy time series data, fuzzy sets design, and conventional time series data. The control of semantic value data is made possible by the use of fuzzy time series data in stock market prediction, which makes accurate value predictions possible. These techniques are mostly used to forecast dynamic, non-linear datasets in a variety of industries, such as stock markets and cryptocurrencies. In order to predict stock prices, artificial intelligence (AI) methods such as neural networks (NN) have become increasingly important. Many methods have been developed to validate parametric and non-parametric theories for predicting trends and gains in the stock market using the feed-forward (FF) NN technique. A variety of soft computing techniques have been investigated in order to provide ways for decision-making based on profit and loss criteria without relying on artificial intelligence. This paper's main goal is to perform a comprehensive examination of the different prediction techniques used to estimate price and return in the stock market. The study includes a comparative review of several approaches as well as a variety of stock market forecasting strategies. © 2024 IEEE.

关键词： Marketplaces

来源：评论

学校读者我要写书评

暂无评论

Gradient Descent Robustly Learns the Intrinsic Dimension of data in Training Convolutional Neural Networks

arXiv

引用

arXiv 2025年

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

Modern neural networks are usually highly over-parameterized. Behind the wide usage of over-parameterized networks is the belief that, if the data are simple, then the trained network will be automatically equivalent to a simple predictor. Following this intuition, many existing works have studied different notions of "ranks" of neural networks and their relation to the rank of data. In this work, we study the rank of convolutional neural networks (CNNs) trained by gradient descent, with a specific focus on the robustness of the rank to image background noises. Specifically, we point out that, when adding background noises to images, the rank of the CNN trained with gradient descent is affected far less compared with the rank of the data. We support our claim with a theoretical case study, where we consider a particular data model to characterize low-rank clean images with added background noises. We prove that CNNs trained by gradient descent can learn the intrinsic dimension of clean images, despite the presence of relatively large background noises. We also conduct experiments on synthetic and real datasets to further validate our claim. © 2025, CC BY-SA.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Benchmarking data Heterogeneity Evaluation Approaches for Personalized Federated Learning

Benchmarking Data Heterogeneity Evaluation Approaches for P...

引用

International Workshop on Trustworthy Federated Learning, FL 2024

作者： Li, Zhilong Wu, Xiaohu Tang, Xiaoli He, Tiantian Ong, Yew-Soon Chen, Mengmeng Liu, Qiqi Lao, Qicheng Yu, Han Beijing University of Posts and Telecommunications Beijing China College of Computing and Data Science Nanyang Technological University Singapore Singapore Institute of High Performance Computing Agency for Science Technology and Research Singapore Singapore Centre for Frontier AI Research Agency for Science Technology and Research Singapore Singapore School of Engineering Westlake University Hangzhou China

ISBN: (纸本)9783031822391

There is growing research interest in measuring the statistical heterogeneity of clients’ local datasets. Such measurements are used to estimate the suitability for collaborative training of personalized federated learning (PFL) models. Currently, these research endeavors are taking place in silos and there is a lack of a unified benchmark to provide a fair and convenient comparison among various approaches in common settings. We aim to bridge this important gap in this paper. The proposed benchmarking framework currently includes six representative approaches. Extensive experiments have been conducted to compare these approaches under five standard non-IID FL settings, providing much needed insights into which approaches are advantageous under which settings. The proposed framework offers useful guidance on the suitability of various data divergence measures in FL systems. It is beneficial for keeping related research activities on the right track in terms of: (1) designing PFL schemes, (2) selecting appropriate data heterogeneity evaluation approaches for specific FL application scenarios, and (3) addressing fairness issues in collaborative model training. The code is available at https://***/Xiaoni-61/DH-Benchmark. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Explainable AI in Blockchain System for Decentralized Governance

Explainable AI in Blockchain System for Decentralized Govern...

引用

2025 International Conference on Pervasive Computational Technologies, ICPCT 2025

作者： Chahar, Suman Kaur, Kawaljit Kaswan, Kuldeep Singh Dhatterwal, Jagjit Singh University Institute of Computing Chandigarh University Mohali India Galgotias University Department of Computer Science and Engineering Greater Noida India Koneru Lakshmaiah Education Foundation Department of Artificial Intelligence and Data Science Andhra Pradesh India

ISBN: (纸本)9798331508685

Decentralized governance is going through radical changes due to the innovative use of blockchain which allows decisions to be made in a transparent, secure and more im- portantly, immutable way, without the authority of a single individual or party. Thus, the introduction of artificial intelligence (AI) in blockchain governing systems complicates such thing as transparency and trust, especially in the case of deep AI systems. But more on this later - this is where Explainable AI or XAI comes into the picture. XAI methods enable understanding of the reasons behind the choices made by AI models and solves the "black box"problem that is predominant in machine learn- ing. When applied to decentralised governance, XAI provides clarifications of AI's decision making in such areas as voting, resource distribution and the enactment of smart contracts. This way, the stakeholders are confident that the decision-making is fair and impartial and this is made possible by the use of XAI. This paper aims to find out more about the applicability of XAI concept in blockchain -based governance systems in terms of bias and fairness in automated decisions and explanation of the decisions made. The reasons why interpretability plays the crucial role in decentralized systems are explained with the help of the comparison between the classical artificial intelligence and the Xin Intelligent. The study also looks at various use cases such as, Decentralized Autonomous Organizations (DAOs), and voting systems implemented on blockchain platforms and how XAI can enhance the performance of the governance systems. As this study shows, XAI is vital in building trust, improving responsibility, and guaranteeing balanced majorities in decentralized blockchain governance structures. © 2025 IEEE.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Cross-domain NER with Generated Task-Oriented Knowledge: An Empirical Study from Information Density Perspective

Cross-domain NER with Generated Task-Oriented Knowledge: An ...

引用

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Zhang, Zhihao Lee, Sophia Yat Mei Wu, Junshuang Zhang, Dong Li, Shoushan Cambria, Erik Zhou, Guodong School of Computer Science & Technology NLP Lab Soochow University China Department of Chinese and Bilingual Studies The Hong Kong Polytechnic University Hong Kong Beijing Jinghang Research Institute of Computing and Communication China College of Computing and Data Science Nanyang Technological University Singapore

ISBN: (纸本)9798891761643

Cross-domain Named Entity Recognition (CDNER) is crucial for Knowledge Graph (KG) construction and natural language processing (NLP), enabling learning from source to target domains with limited data. Previous studies often rely on manually collected entity-relevant sentences from the web or attempt to bridge the gap between tokens and entity labels across domains. These approaches are time-consuming and inefficient, as these data are often weakly correlated with the target task and require extensive pre-training. To address these issues, we propose automatically generating task-oriented knowledge (GTOK) using large language models (LLMs), focusing on the reasoning process of entity extraction. Then, we employ task-oriented pre-training (TOPT) to facilitate domain adaptation. Additionally, current cross-domain NER methods often lack explicit explanations for their effectiveness. Therefore, we introduce the concept of information density to better evaluate the model's effectiveness before performing entity recognition. We conduct systematic experiments and analyses to demonstrate the effectiveness of our proposed approach and the validity of using information density for model evaluation. © 2024 Association for Computational Linguistics.

关键词： Domain Knowledge

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：