检索结果-内蒙古大学图书馆

40th Conference on Uncertainty in Artificial Intelligence, UAI 2024

作者： Bui, Long Minh Huu, Tho Tran Dinh, Duy Nguyen, Tan Minh Hoang, Trong Nghia FPT Software AI Center Department of Mathematics National University of Singapore Singapore School of Electrical Engineering and Computer Science Washington State University United States

Transformers have increasingly become the de facto method to model sequential data with state-of-the-art performance. Due to its widespread use, being able to estimate and calibrate its modeling uncertainty is important to understand and design robust transformer models. To achieve this, previous works have used Gaussian processes (GPs) to perform uncertainty calibration for the attention units of transformers and attained notable successes. However, such approaches have to confine the transformers to the space of symmetric attention to ensure the necessary symmetric requirement of their GP’s kernel specification, which reduces the representation capacity of the model. To mitigate this restriction, we propose the Correlated Gaussian Process Transformer (CGPT), a new class of transformers whose self-attention units are modeled as cross-covariance between two correlated GPs (CGPs). This allows asymmetries in attention and can enhance the representation capacity of GP-based transformers. We also derive a sparse approximation for CGP to make it scale better. Our empirical studies show that both CGP-based and sparse CGP-based transformers achieve better performance than state-of-the-art GP-based transformers on a variety of benchmark tasks. © 2024 Proceedings of Machine Learning Research. All rights reserved.

关键词： Distribution transformers

来源：评论

学校读者我要写书评

暂无评论

Application of Neural Network Keyword Extraction Methods for Student's CV Compilation from Discipline Work Programs 26

Application of Neural Network Keyword Extraction Methods for...

引用

26th International Conference on Soft Computing and Measurements, SCM 2023

作者： Kholod, Ivan I. Korytov, Pavel V. Sorochina, Maria V. Saint-Petersburg Electrotechnical University LETI Department of Computer Science and Engineering Saint-Petersburg Russia Saint-Petersburg Electrotechnical University LETI Department of Software Engineering and Computer Applications Saint-Petersburg Russia

ISBN: (纸本)9798350322484

The paper considers application of natural language processing methods (NLP) to solve the task of identifying the core skills acquired by students in the process of education. An ensemble solution, consisting of models WikiNEuRAL, KBIR-Inspec, ruGTP3Large, word2vec, and other NLP methods, is used to extract and process Russian language keywords. The described method was developed to automate the compilation of a student's CV in the information system called 'Partner's Account' of the LETI' university. © 2023 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning

arXiv

引用

arXiv 2025年

作者： Gu, Wenhao Gu, Li Wang, Ziqiang Suen, Ching Yee Wang, Yang Department of Computer Science and Software Engineering Concordia University Canada

Despite recent significant advancements in Handwritten Document Recognition (HDR), the efficient and accurate recognition of text against complex backgrounds, diverse handwriting styles, and varying document layouts remains a practical challenge. Moreover, this issue is seldom addressed in academic research, particularly in scenarios with minimal annotated data available. In this paper, we introduce the DocTTT framework to address these challenges. The key innovation of our approach is that it uses test-time training to adapt the model to each specific input during testing. We propose a novel Meta-Auxiliary learning approach that combines Meta-learning and self-supervised Masked Autoencoder (MAE). During testing, we adapt the visual representation parameters using a self-supervised MAE loss. During training, we learn the model parameters using a meta-learning framework, so that the model parameters are learned to adapt to a new input effectively. Experimental results show that our proposed method significantly outperforms existing state-of-the-art approaches on benchmark datasets. Copyright © 2025, The Authors. All rights reserved.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

A Systematic Review of Approaches for Reviewing Security-Related Aspects in Agile Requirements Specification of Web Applications

A Systematic Review of Approaches for Reviewing Security-Rel...

引用

2023 Congress in computer science, computer engineering, and Applied Computing, CSCE 2023

作者： Munir, Aqsa Assiri, Mohammed Alam, Saba Naveed Khan, Muzna Butt, Wasi Haider Humayun, Mamoona Department of Computer and Software Engineering Islamabad Pakistan College of Sciences and Humanities Aflaj Prince Sattam Bin Abdulaziz University Department of Computer Science Saudi Arabia College of Computer Engineering and Sciences Prince Sattam Bin Abdulaziz University Department of Software Engineering Saudi Arabia College of Computer and Information Sciences Jouf University Department of Information Systems Al-Jouf Saudi Arabia

ISBN: (纸本)9798350327595

Imperfections in requirement specification can cause serious issues during software development life cycle. It might bring about inferior quality products due to missing attributes, for example, security. Specifically, Web Applications are considered as obvious target for getting significant information. Security aspects have gotten hard to manage in web applications because security requirements are not regularly seen appropriately and often details are missing which lead into ill-characterized security-related aspects. With the help of systemic literature review (SLR), we have identified 32 major research works which were published in period of year 2010 to 2020. We identified 38 security-related aspects of web applications and also 13 techniques and tools for security related aspects in agile requirement specification. We have also analyzed 22 challenges in reviewing security related aspects in agile development life cycle. © 2023 IEEE.

关键词： Agile software development Quality requirements Requirement engineering Security specification Web applications

来源：评论

学校读者我要写书评

暂无评论

Link-Privacy Preserving Graph Embedding Data Publication with Adversarial Learning

引用

Tsinghua science and Technology 2022年第2期27卷 244-256页

作者： Kainan Zhang Zhi Tian Zhipeng Cai Daehee Seo the Department of Computer Science Georgia State UniversityAtlantaGA 30303USA the Department of Electrical and Computer Engineering George Mason UniversityFairfaxVA 22030USA the National Center of Excellence in Software Sangmyung UniversitySeoul 03016Republic of Korea

The inefficient utilization of ubiquitous graph data with combinatorial structures necessitates graph embedding methods,aiming at learning a continuous vector space for the graph,which is amenable to be adopted in traditional machine learning algorithms in favor of vector *** embedding methods build an important bridge between social network analysis and data analytics,as social networks naturally generate an unprecedented volume of graph data *** social network data not only brings benefit for public health,disaster response,commercial promotion,and many other applications,but also gives birth to threats that jeopardize each individual’s privacy and ***,most existing works in publishing social graph embedding data only focus on preserving social graph structure with less attention paid to the privacy issues inherited from social *** be specific,attackers can infer the presence of a sensitive relationship between two individuals by training a predictive model with the exposed social network *** this paper,we propose a novel link-privacy preserved graph embedding framework using adversarial learning,which can reduce adversary’s prediction accuracy on sensitive links,while persevering sufficient non-sensitive information,such as graph topology and node attributes in graph *** experiments are conducted to evaluate the proposed framework using ground truth social network datasets.

关键词： graph embedding privacy preservation adversarial learning

来源：评论

学校读者我要写书评

暂无评论

Explainable AI for Healthcare: An Approach Towards Interpretable Healthcare Models 24

Explainable AI for Healthcare: An Approach Towards Interpret...

引用

24th International Arab Conference on Information Technology, ACIT 2023

作者： Zahoor, Kanwal Bawany, Narmeen Zakaria Ghani, Usman Jinnah University for Women Institute of Business Administration Center for Computing Research Department of Computer Science and Software Engineering Karachi Pakistan

ISBN: (纸本)9798350384307

Artificial intelligence (AI) along with deep learning techniques has become an integral part of almost all aspects of life. One of the domains significantly impacted by this technological revolution is healthcare. Deep learning-based AI systems assist clinicians and medical professionals in disease diagnosis, personalized treatment, and monitoring through wearables, among other applications. Despite its expedient integration in healthcare, the trustworthiness of deep learning models remains a concern, primarily due to a lack of understanding of their underlying processes. However, Explainable AI (XAI) offers explanations through various methods, including Local Interpretable Model-agnostic Explanations (LIME), Shapley Additive explanation (SHAP), and GRAD-CAM. XAI is utilized to enhance transparency, allowing users to understand and trust AI decisions. In this study, we present deep learning models for the classification of pneumonia disease in Chest X-ray Images followed by their explanations. Convolutional Neural Networks (CNNs) and other pre-trained models, including VGG16, MobileNetV3, and ResNet50, were used for classification of images as 'normal' or 'pneumonia'. The VGG16 model, known for its exceptional image understanding capabilities, achieved the highest accuracy, with an impressive 93% score. Further, we used XAI techniques including SHAP, LIME, and Grad-CAM for explanation of models. LIME and Grad-CAM provided more accurate results than SHAP in our experiments. This approach was taken to evaluate the fairness and transparency of the model. The insights gained from XAI can be used to refine and improve machine learning models by identifying areas of weakness or misinterpretation which increases overall model robustness. © 2023 IEEE.

关键词： Deep learning Explainable AI Grad-CAM Healthcare LIME SHAP Transfer learning

来源：评论

学校读者我要写书评

暂无评论

FEW-SHOT LEARNING AS CLUSTER-INDUCED VORONOI DIAGRAMS: A GEOMETRIC APPROACH 10

FEW-SHOT LEARNING AS CLUSTER-INDUCED VORONOI DIAGRAMS: A GEO...

引用

10th International Conference on Learning Representations, ICLR 2022

作者： Ma, Chunwei Huang, Ziyun Gao, Mingchen Xu, Jinhui Department of Computer Science and Engineering University at Buffalo United States Computer Science and Software Engineering Penn State Erie United States

Few-shot learning (FSL) is the process of rapid generalization from abundant base samples to inadequate novel samples. Despite extensive research in recent years, FSL is still not yet able to generate satisfactory solutions for a wide range of real-world applications. To confront this challenge, we study the FSL problem from a geometric point of view in this paper. One observation is that the widely embraced ProtoNet model is essentially a Voronoi Diagram (VD) in the feature space. We retrofit it by making use of a recent advance in computational geometry called Cluster-induced Voronoi Diagram (CIVD). Starting from the simplest nearest neighbor model, CIVD gradually incorporates cluster-to-point and then cluster-to-cluster relationships for space subdivision, which is used to improve the accuracy and robustness at multiple stages of FSL. Specifically, we use CIVD (1) to integrate parametric and nonparametric few-shot classifiers;(2) to combine feature representation and surrogate representation;(3) and to leverage feature-level, transformation-level, and geometry-level heterogeneities for a better ensemble. Our CIVD-based workflow enables us to achieve new state-of-the-art results on mini-ImageNet, CUB, and tiered-ImagenNet datasets, with ∼2%-5% improvements upon the next best. To summarize, CIVD provides a mathematically elegant and geometrically interpretable framework that compensates for extreme data insufficiency, prevents overfitting, and allows for fast geometric ensemble for thousands of individual VD. These together make FSL stronger. © 2022 ICLR 2022 - 10th International Conference on Learning Representationss. All rights reserved.

关键词： Computational geometry

来源：评论

学校读者我要写书评

暂无评论

Dispatching Shots Among Multiple Quantum computers: An Architectural Proposal 4

Dispatching Shots Among Multiple Quantum Computers: An Archi...

引用

4th IEEE International Conference on Quantum Computing and engineering, QCE 2023

作者： Bisicchia, Giuseppe Garcia-Alonso, Jose Murillo, Juan M. Brogi, Antonio University of Pisa Department of Computer Science Pisa Italy University of Extremadura Quercus Software Engineering Group Cáceres Spain

ISBN: (纸本)9798350343236

Quantum Computing is continuously evolving and expanding. As time goes by, more and more Quantum computer implementations become available, each of them with their own features. In such a scenario, it can be difficult for developers to identify which Quantum computer is the most suitable for their needs. In this paper, different from current works presenting strategies to select only one Quantum computer, we propose a change of perspective. Indeed, due to the probabilistic nature of Quantum Mechanics, performing a computation in a Quantum computer usually requires iterating the same execution many times (called shots), to eventually end with a distribution of the final results. Leveraging this need, our architecture enables selecting many Quantum computers for the same circuit and spreading the shots among them. Such a mechanism offers also the possibility for developers to access the partial distributions obtained from the output of a subset of the selected computers. Finally, our architecture proposes to decouple the decision process from the actual execution of such a decision, by enabling developers to encode their specific custom policies. © 2023 IEEE.

关键词： software engineering

来源：评论

学校读者我要写书评

暂无评论

GPT-Based Models Meet Simulation: How to Efficiently use Large-Scale Pre-Trained Language Models Across Simulation Tasks

GPT-Based Models Meet Simulation: How to Efficiently use Lar...

引用

2023 Winter Simulation Conference, WSC 2023

作者： Giabbanelli, Philippe J. Miami University Department of Computer Science & Software Engineering 501 East High Street OxfordOH45056 United States

ISBN: (纸本)9798350369663

The disruptive technology provided by large-scale pre-trained language models (LLMs) such as ChatGPT or GPT-4 has received significant attention in several application domains, often with an emphasis on high-level opportunities and concerns. This paper is the first examination regarding the use of LLMs for scientific simulations. We focus on four modeling and simulation tasks, each time assessing the expected benefits and limitations of LLMs while providing practical guidance for modelers regarding the steps involved. The first task is devoted to explaining the structure of a conceptual model to promote the engagement of participants in the modeling process. The second task focuses on summarizing simulation outputs, so that model users can identify a preferred scenario. The third task seeks to broaden accessibility to simulation platforms by conveying the insights of simulation visualizations via text. Finally, the last task evokes the possibility of explaining simulation errors and providing guidance to resolve them. © 2023 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Breast Microcalcification detection in digital mammograms using Deep Transfer learning approaches 23

Breast Microcalcification detection in digital mammograms us...

引用

9th International Conference on Computing and Data engineering, ICCDE 2023

作者： Rasool, Ehtsham Anwar, Muhammad Junaid Shaker, Bilawal Hashmi, Muhammad Harris Rehman, Khalil Ur Seed, Yousaf Department of Computer Science The University of Alabama Birmingham United States School of Software Engineering Beijing University of Technology Beijing China

ISBN: (纸本)9781450398022

Breast cancer is the most often diagnosed cancer in women affecting one in eight at the age of 80 in US. Breast is the most threatening cancer among women which leads to death. Early diagnosis of breast cancer can save their lives which decreases the mortality rate. Mammography is a standard screening method for breast cancer diagnosis that identifies occurrences of breast cancer in women's at early stages without symptoms. In this study, we employed transfer learning in deep learning to increase the neural network's performance and reduce the false positive rate. In addition, we proposed a pre-trained VGG-19 neural network to extract features of individual microcalcification to predict breast cancer. The proposed method was evaluated on two public databases the CBIS-DDSM and DDSM and achieved 0.98 sensitivities respectively. The proposed method obtained higher sensitivity than other residual neural networks and previous studies. © 2023 ACM.

关键词： Diseases

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：