检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Yang, Yunhao Xue, Bo Hao, Yunzhi Li, Ying Wan, Yuanyu School of Software Technology Zhejiang University Ningbo China Department of Computer Science City University of Hong Kong China Big Graph Center School of Computer and Computing Science Hangzhou City University Hangzhou China State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou China Bangsheng Technology Co. Ltd. Hangzhou China Institute of Blockchain and Data Security Hangzhou China

ISBN: (纸本)157735897X

We investigate safe online convex optimization (SOCO), where each decision must satisfy a set of unknown linear constraints. Assuming that the unknown constraints can be observed with a sub-Gaussian noise for each chosen decision, previous studies have established a high-probability regret bound of O(T2/3). However, this assumption may not hold in many practical scenarios. To address this limitation, in this paper, we relax the assumption to allow any noise that admits finite (1+ϵ)-th moments for some ϵ ∈ (0, 1], and propose two algorithms that enjoy an O(Tcϵ) regret bound with high probability, where T is the time horizon and cϵ = (1+ϵ)/(1+2ϵ). The key idea of our two algorithms is to respectively utilize the median-of-means and truncation techniques to achieve accurate estimation under heavy-tailed noises. To the best of our knowledge, these are the first algorithms designed to handle SOCO with heavy-tailed observation noises. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Convex optimization

来源：评论

学校读者我要写书评

暂无评论

Last-X-Generation Archiving Strategy for Multi-Objective Evolutionary Algorithms 13

Last-X-Generation Archiving Strategy for Multi-Objective Evo...

引用

13th IEEE Congress on Evolutionary Computation, CEC 2024

作者： Shu, Tianye Nan, Yang Shang, Ke Ishibuchi, Hisao Southern University of Science and Technology Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Shenzhen518055 China Southern University of Science and Technology Shenzhen518055 China Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen518060 China

ISBN: (纸本)9798350308365

For evolutionary multi-objective optimization algorithms (EMOAs), an external archive can be utilized for saving good solutions found throughout the evolutionary process. Recent studies showed that a solution set selected from an external archive is usually superior to the final population. That is, the incorporation of an external archive improves the performance of EMOAs. However, the computation time for maintaining an external archive is long, especially when the archive size is large. To solve this issue, a simple archiving strategy is to save all solutions generated in the last several generations. In this paper, we examine this archiving strategy for three representative EMOAs on artificial test problems (Minus-DTLZ and WFG) and real-world problems (RE). Our results show that archiving the last several generations clearly improves the performance of EMOAs without severely increasing the computation time. © 2024 IEEE.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Transforming the Client Relationship to Support Large Capstone Classes

Transforming the Client Relationship to Support Large Capsto...

引用

Frontiers in Education (FIE) Conference

作者： Bowen Hui Samantha Hodge Dilpreet Samra Department of Computer Science Mathematics Physics and Statistics University of British Columbia Kelowna Canada

ISBN: (数字)9798350351507

ISBN: (纸本)9798350363067

Ahstract–This innovative practice full paper describes a case study from a software engineering capstone project course. Undergraduate programs often have a final-year capstone course designed to integrate and apply the knowledge and skills students have previously acquired while adapting to industry-standard practices. Capstones play a critical role in bridging the gap between academia and the industry as students transition to the workforce. Over the past twelve years, our institution has adopted a client-based model where industry clients work closely with a single team to solve a real-world problem. However, rising enrollment has put a strain on running this model effectively because of difficulties in recruiting clients, managing numerous client relationships simultaneously, and keeping client-student interactions sustainable. To tackle these challenges, we propose a new client model where clients pitch their ideas as themes in a competition and act as panel judges in evaluating student team submissions. We call this the hackathon client model and evaluate it in a class with 22 teams and 104 students. Through a thematic analysis of the qualitative responses gathered from this study, our findings suggest this new model provides a scalable alternative to operating a large capstone class while preserving many of the benefits of the traditional client model. However, both students and clients indicate having more means of communication would improve the project requirements phase and strengthen their relationship. We discuss ideas on improving the hackathon client model and plans for future experimentation in large capstones.

关键词： Industries Analytical models Adaptation models Computational modeling Problem-solving Hackathon Programming profession Strain Software engineering Logistics

来源：评论

学校读者我要写书评

暂无评论

Self-supervised Benchmark Lottery on ImageNet: Do Marginal Improvements Translate to Improvements on Similar datasets?

Self-supervised Benchmark Lottery on ImageNet: Do Marginal I...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Utku Ozbulak Esla Timothy Anzaku Solha Kang Wesley De Neve Joris Vankerschaver Center for Biosystems and Biotech Data Science Ghent University Global Campus Republic of Korea Department of Electronics and Information Systems Ghent University Belgium Department of Applied Mathematics Computer Science and Statistics Ghent University Belgium

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Machine learning (ML) research strongly relies on benchmarks in order to determine the relative effectiveness of newly proposed models. Recently, a number of prominent research effort argued that a number of models that improve the state-of-the-art by a small margin tend to do so by winning what they call a "benchmark lottery". An important benchmark in the field of machine learning and computer vision is the ImageNet where newly proposed models are often showcased based on their performance on this dataset. Given the large number of self-supervised learning (SSL) frameworks that has been proposed in the past couple of years each coming with marginal improvements on the ImageNet dataset, in this work, we evaluate whether those marginal improvements on ImageNet translate to improvements on similar datasets or not. To do so, we investigate twelve popular SSL frameworks on five ImageNet variants and discover that models that seem to perform well on ImageNet may experience significant performance declines on similar datasets. Specifically, state-of-the-art frameworks such as DINO and Swav, which are praised for their performance, exhibit substantial drops in performance while MoCo and Barlow Twins displays comparatively good results. As a result, we argue that otherwise good and desirable properties of models remain hidden when benchmarking is only performed on the ImageNet validation set, making us call for more adequate benchmarking. To avoid the "benchmark lottery" on ImageNet and to ensure a fair benchmarking process, we investigate the usage of a unified metric that takes into account the performance of models on other ImageNet variant datasets.

关键词： Measurement computer vision Computational modeling Neural networks Machine learning Self-supervised learning Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

Fine-grained Text-Video Fusion for Referring Video Object Segmentation

Fine-grained Text-Video Fusion for Referring Video Object Se...

引用

Power Electronics, computer Applications (ICPECA), IEEE International Conference on

作者： Yonglin Li Jiaqi Wan Xiao Teng Long Lan Department of Intelligent Data Science College of Computer Science and Technology National University of Defense Technology Changsha China

Referring video object segmentation (RVOS) aims at segmenting an object in a video with its text description. The core of RVOS lies in the modal alignment between the vision and text. To improve the performance, most previous RVOS methods are devoted to exploring more sophisticated visual cues. However, these methods fail to fully exert the inherent structured properties of text and only use coarse text features to tackle vision-text interaction. In this paper, we propose FTVR, an approach that utilizes Fine-grain Text-Video method for Referring video object segmentation to explore fine-grain text information. We introduce a LLM to split the text sentence into different functional phrases and propose three novel modules to enhance the cross-modal alignment. Concretely, we design a dynamic-aware perception module to deal with motion-related phrases and a following global-aware attention module to fuse the outputted motion information. To deal with entity-related phrases, our FTVR also introduces an entity-aware augmentation module to highlight entity information. Extensive experiments show the effectiveness of our method on four popular benchmarks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Comparison of the SOLA model and ARIMA for Cloud Computing Resources Allocations

Comparison of the SOLA model and ARIMA for Cloud Computing R...

引用

International Multidisciplinary Information Technology and Engineering Conference (IMITEC)

作者： Sello Prince Sekwatlakwatla Vusumuzi Malele Department of Computer Science and Information Systems Unit for Data Science and Computing North-West University South Africa

ISBN: (数字)9798350387988

ISBN: (纸本)9798350387995

Cloud computing solutions are becoming more and more popular as a way for organizations to improve productivity, save costs, and simplify procedures. The advantage of cloud services is that they enable users to store data remotely and access on-demand applications and services from a shared pool of configurable computing resources. Cloud computing resource allocation faces challenges in load balancing, effective resource management, and compliance with regulatory and legal requirements, requiring complex tools and techniques for optimal use. Most organizations are moving to cloud services and encouraging their clients to access services online. However, effective resource allocation is critical for improving performance and lowering costs in this area. Due to unpredictable network traffic in cloud computing, resource allocation is challenging, which causes customers to complain about application timeouts, delayed system times, and higher bandwidth use during peak hours. This entails allocating resources to various users and programs, such as memory, processing power, storage, and network bandwidth. In this regard, this study compares the ensemble method, which is stepwise Gaussian Linear Autoregressive (SGLA) and the individual method, which is autoregressive integrated moving average (ARIMA). The results show SGLA prediction accuracy increased with an average of 97.9%, and ARIMA prediction showed an accuracy of 71.5%. In this regard the ensemble method performed better than individual methods using the same datasets. The study recommends the ensemble method for the prediction and allocation of resources in cloud computing.

关键词： Productivity Cloud computing Accuracy Costs Organizations Bandwidth Telecommunication traffic Resource management Ensemble learning Servers

来源：评论

学校读者我要写书评

暂无评论

On Newton’s Method to Unlearn Neural Networks

arXiv

引用

arXiv 2024年

作者： Nhung, Bui Thi Cam Lu, Xinyang Sim, Rachael Hwee Ling Ng, See-Kiong Low, Bryan Kian Hsiang Department of Computer Science National University of Singapore Singapore Institute of Data Science National University of Singapore Singapore

With the widespread applications of neural networks (NNs) trained on personal data, machine unlearning has become increasingly important for enabling individuals to exercise their personal data ownership, particularly the "right to be forgotten" from trained NNs. Since retraining is computationally expensive, we seek approximate unlearning algorithms for NNs that return identical models to the retrained oracle. While Newton’s method has been successfully used to approximately unlearn linear models, we observe that adapting it for NN is challenging due to degenerate Hessians that make computing Newton’s update impossible. Additionally, we show that when coupled with popular techniques to resolve the degeneracy, Newton’s method often incurs offensively large norm updates and empirically degrades model performance post-unlearning. To address these challenges, we propose CureNewton’s method, a principle approach that leverages cubic regularization to handle the Hessian degeneracy effectively. The added regularizer eliminates the need for manual finetuning and affords a natural interpretation within the unlearning context. Experiments across different models and datasets show that our method can achieve competitive unlearning performance to the state-of-the-art algorithm in practical unlearning settings, while being theoretically justified and efficient in running time. © 2024, CC BY.

关键词： Newton-Raphson method

来源：评论

学校读者我要写书评

暂无评论

A computational supervised neural network procedure for the fractional SIQ mathematical model

引用

European Physical Journal: Special Topics 2023年第5期232卷 535-546页

作者： Mukdasai, Kanit Sabir, Zulqurnain Raja, Muhammad Asif Zahoor Singkibud, Peerapongpat Sadat, R. Ali, Mohamed R. Department of Mathematics Faculty of Science Khon Kaen University Khon Kaen 40002 Thailand Department of Mathematics and Statistics Hazara University Mansehra Pakistan Future Technology Research Center National Yunlin University of Science and Technology 123 University Road Section 3 Yunlin Douliou 64002 Taiwan Department of Applied Mathematics and Statistics Faculty of Science and Liberal Arts Rajamangala University of Technology Isan Nakhon Ratchasima 30000 Thailand Department of Mathematics Faculty of Engineering Zagazig University Zagazig Egypt Faculty of Engineering and Technology Future University in Egypt New Cairo 11835 Egypt Basic Engineering Science Department Benha Faculty of Engineering Benha University Benha Egypt Department of Computer Science and Mathematics Lebanese American University Beirut Lebanon

The purpose of the current work is to provide the numerical solutions of the fractional mathematical system of the susceptible, infected and quarantine (SIQ) system based on the lockdown effects of the coronavirus disease. These investigations provide more accurateness by using the fractional SIQ system. The investigations based on the nonlinear, integer and mathematical form of the SIQ model together with the effects of lockdown are also presented in this work. The impact of the lockdown is classified into the susceptible/infection/quarantine categories, which is based on the system of differential models. The fractional study is provided to find the accurate as well as realistic solutions of the SIQ model using the artificial intelligence (AI) performances along with the scale conjugate gradient (SCG) design, i.e., AI-SCG. The fractional-order derivatives have been used to solve three different cases of the nonlinear SIQ differential model. The statics to perform the numerical results of the fractional SIQ dynamical system are 7% for validation, 82% for training and 11% for testing. To observe the exactness of the AI-SCG procedure, the comparison of the numerical attained performances of the results is presented with the reference Adam solutions. For the validation, authentication, aptitude, consistency and validity of the AI-SCG solver, the computing numerical results have been provided based on the error histograms, state transition measures, correlation/regression values and mean square error. © 2022, The Author(s), under exclusive licence to EDP sciences, Springer-Verlag GmbH Germany, part of Springer Nature.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Anytime-valid FDR control with the stopped e-BH procedure

arXiv

引用

arXiv 2025年

作者： Wang, Hongjian Dandapanthula, Sanjit Ramdas, Aaditya Department of Statistics and Data Science Carnegie Mellon University United States Machine Learning Department Carnegie Mellon University United States

The recent e-Benjamini-Hochberg (e-BH) procedure for multiple hypothesis testing is known to control the false discovery rate (FDR) under arbitrary dependence between the input e-values. This paper points out an important subtlety when applying the e-BH procedure with e-processes, which are sequential generalizations of e-values (where the data are observed sequentially). Since adaptively stopped e-processes are e-values, the e-BH procedure can be repeatedly applied at every time step, and one can continuously monitor the e-processes and the rejection sets obtained. One would hope that the "stopped e-BH procedure" (se-BH) has an FDR guarantee for the rejection set obtained at any stopping time. However, while this is true if the data in different streams are independent, it is not true in full generality, because each stopped e-process is an e-value only for stopping times in its own local filtration, but the se-BH procedure employs a stopping time with respect to a global filtration. This can cause information to leak across time, allowing one stream to know its future by knowing past data of another stream. This paper formulates a simple causal condition under which local e-processes are also global e-processes and thus the se-BH procedure does indeed control the FDR. The condition excludes unobserved confounding from the past and is met under most reasonable scenarios including genomics. Copyright © 2025, The Authors. All rights reserved.

关键词： Filtration

来源：评论

学校读者我要写书评

暂无评论

Recognizing Sentiments In Regional Language Text From Social Media Using Machine Learning Approach: A comparative Study to understand the state of the art tools and associated functionalities available 3

Recognizing Sentiments In Regional Language Text From Social...

引用

3rd International Conference on Advanced Computing Technologies and Applications, ICACTA 2023

作者： Sirsat, Sudarshan Zulpe, Nitish K J Somaiya Institute of Management Somaiaya Vidyavihar University Department. of Data Science and Technology Mumbai India Swami Ramanand Teerth Marathwada Univesity College of Computer Science and Information Technology Latur India

ISBN: (纸本)9798350348347

Regional language contents gain the highest traction in today's social media network marketing and from users' perspectives since nearly 90% of users consume contents in regional language. Those who are only focusing on English as the major language of social media or website contents are clearly missing the majority. Social media currently is touching the untouched digital population of the world through creative contents, easy to use convincing features and infinite possibilities of enhancing personal social networks using regional language like Hindi, Marathi, and Gujarati and many other state and regional languages. This has created immense opportunities in the area of text mining, text analysis and text generation too many scholars and researchers to visualize and analyses the contents to gain insights into what are the multidimensional perspectives of the authors of any such contents. This research work is going to understand the current work done in the area of regional language sentiment analysis in India and all across the globe to understand the status and research gap in the domain to work further in the Maharashtra state regional language contents available on social media pages. This research work will try to explore all possible tools and technologies available and used by the recent research scholars and industry leaders to decide the way ahead to work in the Marathi regional language domain in near future. © 2023 IEEE.

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：