检索结果-内蒙古大学图书馆

SSRN 2023年

作者： Wang, Guohong Ma, Hua Gao, Yansong Abuadbba, Alsharif Zhang, Zhi Wei, Kang Al-Sarawi, Said F. Zhang, Gongxuan Abbott, Derek School of Cyber Science and Engineering Nanjing University of Science and Technology Jiangsu Nanjing China School of Electrical and Electronic Engineering The University of Adelaide Australia Department of Computer Science and Software Engineering The University of Western Australia Australia Data61 Commonwealth Scientific and Industrial Research Organisation CSIRO Australia

Image camouflage has been utilized to create clean-label poisoned images for implanting backdoor into a DL model. But there exists a crucial limitation that one attack/poisoned image can only fit a single input size of the DL model, which greatly increases its attack budget when attacking multiple commonly adopted input sizes of DL models. This work proposes to constructively craft an attack image through camouflaging but can fit multiple DL models’ input sizes simultaneously, namely OmClic. Thus, through OmClic, we are able to always implant a backdoor regardless of which common input size is chosen by the user to train the DL model given the same attack budget (i.e., a fraction of the poisoning rate). With our camouflaging algorithm formulated as a multi-objective optimization, M = 5 input sizes can be concurrently targeted with one attack image, which artifact is retained to be almost visually imperceptible at the same time. Extensive evaluations validate the proposed OmClic can reliably succeed in various settings using diverse types of images. Further experiments on OmClic based backdoor insertion to DL models show that high backdoor performances (i.e., attack success rate and clean data accuracy) are achievable no matter which common input size is randomly chosen by the user to train the model. So that the OmClic based backdoor attack budget is reduced by M× compared to the state-of-the-art camouflage based backdoor attack as a baseline. Significantly, the same set of OmClic based poisonous attack images is transferable to different model architectures for backdoor implant. © 2023, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

ICEBAR: Feedback-Driven Iterative Repair of Alloy Specifications 22

ICEBAR: Feedback-Driven Iterative Repair of Alloy Specificat...

引用

Proceedings of the 37th IEEE/ACM International Conference on Automated software engineering

作者： Simón Gutiérrez Brida Germán Regis Guolong Zheng Hamid Bagheri Thanhvu Nguyen Nazareno Aguirre Marcelo Frias Department of Computer Science University of Rio Cuarto Argentina and CONICET Argentina Department of Computer Science University of Rio Cuarto Argentina University of Nebraska-Lincoln United States of America George Mason University United States of America Department of Software Engineering Buenos Aires Institute of Technology Argentina and CONICET Argentina

ISBN: (纸本)9781450394758

Automated program repair (APR) techniques have shown great success in automatically finding fixes for programs in programming languages such as C or Java. In this work, we focus on repairing formal specifications, in particular for the Alloy specification language. As opposed to most APR tools, our approach to repair Alloy specifications, named ICEBAR, does not use test-based oracles for patch assessment. Instead, ICEBAR relies on the use of property-based oracles, commonly found in Alloy specifications as predicates and assertions. These property-based oracles define stronger conditions for patch assessment, thus reducing the notorious overfitting issue caused by using test-based oracles, typically observed in APR contexts. Moreover, as assertions and predicates are inherent to Alloy, whereas test cases are not, our tool is potentially more appealing to Alloy users than test-based Alloy repair tools. At a high level, ICEBAR is an iterative, counterexample-based process, that generates and validates repair candidates. ICEBAR receives a faulty Alloy specification with a failing property-based oracle, and uses Alloy’s counterexamples to build tests and feed ARepair, a test-based Alloy repair tool, in order to produce a repair candidate. The candidate is then checked against the property oracle for overfitting: if the candidate passes, a repair has been found; if not, further counterexamples are generated to construct tests and enhance the test suite, and the process is iterated. ICEBAR includes different mechanisms, with different degrees of reliability, to generate counterexamples from failing predicates and assertions. Our evaluation shows that ICEBAR significantly improves over ARepair, in both reducing overfitting and improving the repair rate. Moreover, ICEBAR shows that iterative refinement allows us to significantly improve a state-of-the-art tool for automated repair of Alloy specifications without any modifications to the tool.

关键词： Formal Specification Alloy Automated Repair

来源：评论

学校读者我要写书评

暂无评论

Federated Learning for Diffusion Models

引用

IEEE Transactions on Cognitive Communications and Networking 2025年

作者： Peng, Zihao Wang, Xijun Chen, Shengbo Rao, Hong Shen, Cong The School of Mathematics and Computer Science Nanchang University Nanchang330000 China The School of Electronics and Information Technology Sun Yat-sen University Guangzhou510006 China The School of Software Nanchang University Nanchang330000 China The Charles L. Brown Department of Electrical and Computer Engineering University of Virginia CharlottesvilleVA22904 United States

Diffusion models are powerful generative models that can produce highly realistic samples for various tasks. Typically, these models are constructed using centralized, independently and identically distributed (IID) training data. However, in practical scenarios, data is often distributed across multiple clients and frequently manifests non-IID characteristics. Federated Learning (FL) can leverage this distributed data to train diffusion models, but the performance of existing FL methods is unsatisfactory in non-IID scenarios. To address this, we propose FEDDDPM—Federated Learning with Denoising Diffusion Probabilistic Models, which leverages the data generative capability of diffusion models to facilitate model training. In particular, the server uses well-trained local diffusion models uploaded by each client before FL training to generate auxiliary data that can approximately represent the global data distribution. Following each round of model aggregation, the server further optimizes the global model using the auxiliary dataset to alleviate the impact of heterogeneous data on model performance. We provide a rigorous convergence analysis of FEDDDPM and propose an enhanced algorithm, FEDDDPM+, to reduce training overheads. FEDDDPM+ detects instances of slow model learning and performs a one-shot correction using the auxiliary dataset. Experimental results validate that our proposed algorithms outperform the state-of-the-art FL algorithms on the MNIST, CIFAR10 and CIFAR100 datasets. © 2025 IEEE. All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Robust Generalization Against Photon-Limited Corruptions via Worst-Case Sharpness Minimization

Robust Generalization Against Photon-Limited Corruptions via...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Zhuo Huang Miaoxi Zhu Xiaobo Xia Li Shen Jun Yu Chen Gong Bo Han Bo Du Tongliang Liu Sydney AI Centre The University of Sydney National Engineering Research Center for Multimedia Software Institute of Artificial Intelligence School of Computer Science Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University JD Explore Academy Department of Automation University of Science and Technology of China Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education School of Computer Science and Engineering Nanjing University of Science and Technology Department of Computer Science Hong Kong Baptist University

Robust generalization aims to tackle the most challenging data distributions which are rare in the training set and contain severe noises, i.e., photon-limited corruptions. Common solutions such as distributionally robust optimization (DRO) focus on the worst-case empirical risk to ensure low training error on the uncommon noisy distributions. However, due to the over-parameterized model being optimized on scarce worst-case data, DRO fails to produce a smooth loss landscape, thus struggling on generalizing well to the test set. Therefore, instead of focusing on the worst-case risk minimization, we propose SharpDRO by penalizing the sharpness of the worst-case distribution, which measures the loss changes around the neighbor of learning parameters. Through worst-case sharpness minimization, the proposed method successfully produces a flat loss curve on the corrupted distributions, thus achieving robust generalization. Moreover, by considering whether the distribution annotation is available, we apply SharpDRO to two problem settings and design a worst-case selection process for robust generalization. Theoretically, we show that SharpDRO has a great convergence guarantee. Experimentally, we simulate photon-limited corruptions using CIFAR10/100 and ImageNet30 datasets and show that SharpDRO exhibits a strong generalization ability against severe corruptions and exceeds well-known baseline methods with large performance gains.

关键词：

来源：评论

学校读者我要写书评

暂无评论

TCSloT: Text Guided 3D Context and Slope Aware Triple Network for Dental Implant Position Prediction

arXiv

引用

arXiv 2023年

作者： Yang, Xinquan Xie, Jinheng Li, Xuechen Li, Xuguang Shen, Linlin Deng, Yongqiang College of Computer Science and Software Engineering Shenzhen University Shenzhen China Show Lab National University of Singapore Singapore Department of Stomatology Shenzhen University General Hospital Shenzhen China

In implant prosthesis treatment, the surgical guide of implant is used to ensure accurate implantation. However, such design heavily relies on the manual location of the implant position. When deep neural network has been proposed to assist the dentist in locating the implant position, most of them take a single slice as input, which do not fully explore 3D contextual information and ignoring the influence of implant slope. In this paper, we design a Text Guided 3D Context and Slope Aware Triple Network (TCSloT) which enables the perception of contextual information from multiple adjacent slices and awareness of variation of implant slopes. A Texture Variation Perception (TVP) module is correspondingly elaborated to process the multiple slices and capture the texture variation among slices and a Slope-Aware Loss (SAL) is proposed to dynamically assign varying weights for the regression head. Additionally, we design a conditional text guidance (CTG) module to integrate the text condition (i.e., left, middle and right) from the CLIP for assisting the implant position prediction. Extensive experiments on a dental implant dataset through five-fold cross-validation demonstrated that the proposed TCSloT achieves superior performance than existing methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Random Forest and Multilayer Perceptron Algorithms for Predicting Gross Domestic Product

Comparative Analysis of Random Forest and Multilayer Percept...

引用

International Conference on Emerging Trends in Networks and computer Communications (ETNCC)

作者： Ambrose A. Azeta Muyideen D. Adewale Patrick O. Omote Funmi O. Onaolapo Idowo A. Bakare Steven Tjiraso Department of Software Engineering Namibia University of Science and Technology Namibia. Africa Centre of Excellence on Technology Enhanced Learning National Open University of Nigeria Nigeria. Department of Computer and Mathematical Sciences Crawford University Ogun State Nigeria.

ISBN: (数字)9798350353266

ISBN: (纸本)9798350353273

This study examines the efficacy of machine learning techniques, particularly Multilayer Perceptron (MLP), and Random Forest and in forecasting Nigeria's Gross Domestic Product (GDP). Traditional statistical methods frequently fail to capture intricate, nonlinear relationships, thus necessitating advanced techniques. The dataset employed in this research encompasses 22 instances from 2000 to 2021. It includes variables such as GDP, healthcare expenditure, population, and the index of economic freedom obtained from the World Bank as well as the Nigerian Perceptions Index for corruption. The data was prepared to undergo training and testing following the CRISP-DM methodology. The models were trained and evaluated using a five-member cross-validation method. Performance metrics indicated that MLP outperformed Random Forest, with MAE values of 16.8228 and 25.3074 and RMSE values of 23.2111 and 33.2027, respectively. These findings suggest that MLP offers more accurate GDP predictions for Nigeria, effectively managing complex data relationships. This study provides valuable insights for enhancing predictive accuracy and supporting more effective economic planning and decision-making.

关键词： Training Machine learning algorithms Accuracy Economic indicators Biological system modeling Multilayer perceptrons Predictive models Prediction algorithms Random forests Testing

来源：评论

学校读者我要写书评

暂无评论

Two Birds, One Stone: A Unified Framework for Joint Learning of Image and Video Style Transfers

Two Birds, One Stone: A Unified Framework for Joint Learning...

引用

International Conference on computer Vision (ICCV)

作者： Bohai Gu Heng Fan Libo Zhang Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Department of Computer Science and Engineering University of North Texas Denton TX USA

Current arbitrary style transfer models are limited to either image or video domains. In order to achieve satisfying image and video style transfers, two different models are inevitably required with separate training processes on image and video domains, respectively. In this paper, we show that this can be precluded by introducing UniST, a Unified Style Transfer framework for both images and videos. At the core of UniST is a domain interaction transformer (DIT ), which first explores context information within the specific domain and then interacts contextualized domain information for joint learning. In particular, DIT enables exploration of temporal information from videos for the image style transfer task and meanwhile allows rich appearance texture from images for video style transfer, thus leading to mutual benefits. Considering heavy computation of traditional multi-head self-attention, we present a simple yet effective axial multi-head self-attention (AMSA) for DIT , which improves computational efficiency while maintains style transfer performance. To verify the effectiveness of UniST, we conduct extensive experiments on both image and video style transfer tasks and show that UniST performs favorably against state-of-the-art approaches on both tasks. Code is available at https://***/NevSNev/UniST.

关键词：

来源：评论

学校读者我要写书评

暂无评论

LALISA: Adaptive Bitrate Ladder Optimization in HTTP-based Adaptive Live Streaming

LALISA: Adaptive Bitrate Ladder Optimization in HTTP-based A...

引用

IEEE Symposium on Network Operations and Management

作者： Farzad Tashtarian Abdelhak Bentaleb Hadi Amirpour Babak Taraghi Christian Timmerer Hermann Hellwagner Roger Zimmermann Christian Doppler Laboratory ATHENA Alpen-Adria-Universitat Klagenfurt Austria Department of Computer Science and Software Engineering Concordia University Canada School of Computing National University of Singapore Singapore

Video content in Live HTTP Adaptive Streaming (HAS) is typically encoded using a pre-defined, fixed set of bitrate-resolution pairs (termed Bitrate Ladder), allowing play-back devices to adapt to changing network conditions using an adaptive bitrate (ABR) algorithm. However, using a fixed one-size-fits-all solution when faced with various content complexities, heterogeneous network conditions, viewer device resolutions and locations, does not result in an overall maximal viewer quality of experience (QoE). Here, we consider these factors and design LALISA, an efficient framework for dynamic bitrate ladder optimization in live HAS. LALISA dynamically changes a live video session’s bitrate ladder, allowing improvements in viewer QoE and savings in encoding, storage, and bandwidth costs. LALISA is independent of ABR algorithms and codecs, and is deployed along the path between viewers and the origin server. In particular, it leverages the latest developments in video analytics to collect statistics from video players, content delivery networks and video encoders, to perform bitrate ladder tuning. We evaluate the performance of LALISA against existing solutions in various video streaming scenarios using a trace-driven testbed. Evaluation results demonstrate significant improvements in encoding computation (24.4%) and bandwidth (18.2%) costs with an acceptable QoE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Attribute Based Federated-Reinforcement Learning Approach for Drone Authorization 1

引用

3rd International Conference on Machine Intelligence and Smart Systems, MISS 2023

作者： Rao, K. Rajesh Pradhan, Tribikram Krishna Prakasha, K. Bosch Global Software Technologies Private Ltd. Robert Bosch Bangalore560030 India Department of Computer Science and Engineering Tezpur University Assam Tezpur784028 India Department of Information and Communication Technology Manipal Institute of Technology MAHE Manipal576104 India

ISBN: (数字)9783031317231

ISBN: (纸本)9783031317224

Drones have become a vital part of the technology-driven world we live in today. From using it for delivering our day-to-day items to high grade military purposes, drones have spread their demands to various sectors of the economy. However, a little exploration has been done when it comes to talking about their security. Security generally comprises two aspects: authentication and authorization. While authentication is more to do with confirming the users with who they say they are, authorization is when the user’s permissions to access a resource is examined and a decision to approve or deny the access is taken by the system. This paper discusses authorization in detail using reinforcement learning and federated learning. The model is trained on a synthetic dataset, and is evaluated based on progressive validation loss (PVL), F1-Score, and a new metric called Permit Score. © 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

An Analysis of Depression Detection Techniques from Online Social Networks 2nd

An Analysis of Depression Detection Techniques from Online S...

引用

2nd International Conference on Intelligent Technologies and Applications, INTAP 2019

作者： Bilal, Uffaq Khan, Farhan Hassan Department of Computer Science and Software Engineering IIUI Islamabad Pakistan Islamabad Pakistan

ISBN: (纸本)9789811552311

Mental illness is caused by depression which may have a deep negative impact on individuals or on the society as a whole. It is a growing and severe problem which has the tendency to increase with time due to the extensive use of social networking websites such as Facebook, Twitter, Instagram etc. These social networking websites allow users to share images, videos, expressions and emotions. Depression is a form of mental illness. Patients suffering from depression have mood disorders such as low mood, high mood, lack of interest in things, etc. Machine learning techniques on text data and emojis have been applied to automatically classify a user into depressed and non-depressed. State-of-the-art classifiers have been used by the researcher to detect depressed individuals. Benchmark datasets are composed of text and emojis used in the social networking websites where classification is based on four factors Emotional Process, Temporal Process, Linguistic Style and combination of these three factors. In the proposed method, the emotional process is combined with their respective emojis to develop an automatic system for the detection of depressed patients. The features from the emotional process and emojis will be extracted and state-of-the-art classifiers have been proposed to be trained and evaluated using multiple classifiers using different combinations of part-of-speech tags. © 2020, Springer Nature Singapore Pte Ltd.

关键词： Sentiment analysis

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：