检索结果-内蒙古大学图书馆

Joint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024

作者： Liu, Daizong Qu, Xiaoye Fang, Xiang Dong, Jianfeng Zhou, Pan Nan, Guoshun Tang, Keke Fang, Wanlong Cheng, Yu Wangxuan Institute of Computer Technology Peking University China School of Cyber Science and Engineering Huazhong University of Science of Technology China Interdisciplinary Graduate Programme Nanyang Technological University Singapore College of Computer Science and Technology Zhejiang Gongshang University China School of Cyberspace Security Beijing University of Posts and Telecommunications China Cyberspace Institute of Advanced Technology Guangzhou University China Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong

ISBN: (纸本)9782493814104

This paper addresses the task of temporal activity localization (TAL). Although recent works have made significant progress in TAL research, almost all of them implicitly assume that the dense frame-level correspondences in each video-query pair are correctly annotated. However, in reality, such an assumption is extremely expensive and even impossible to satisfy due to subjective labeling. To alleviate this issue, in this paper, we explore a new TAL setting termed Noisy Temporal activity localization (NTAL), where a TAL model should be robust to the mixed training data with noisy moment boundaries. Inspired by the memorization effect of neural networks, we propose a novel method called Co-Teaching Regularizer (CTR) for NTAL. Specifically, we first learn a Gaussian Mixture Model to divide the mixed training data into preliminary clean and noisy subsets. Subsequently, we refine the labels of the two subsets by an adaptive prediction function so that their true positive and false positive samples could be identified. To avoid single model being prone to its mistakes learned by the mixed data, we adopt a co-teaching paradigm, which utilizes two models sharing the same framework to teach each other for robust learning. A curriculum strategy is further introduced to gradually learn the moment confidence from easy to hard. Experiments on three datasets demonstrate that our CTR is significantly more robust to the noisy training data compared to the existing methods. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

关键词： Gaussian distribution

来源：评论

学校读者我要写书评

暂无评论

Adaptive Droop Control in VSC-Based MT-MVDC Distribution Systems Considering ESS Participation

引用

IEEE Transactions on Energy Conversion 2024年第2期40卷 761-772页

作者： Sun, Pingyang Pinthurat, Watcharakorn Li, Gen Khalid, Muhammad Town, Graham Konstantinou, Georgios UNSW Sydney School of Electrical Engineering and Telecommunications NSW2052 Australia Rajamangala University of Technology Tawan-Ok Department of Electrical Engineering Chanthaburi Thailand Energy Technology and Computer Science Section Department of Engineering Technology Ballerup2750 Denmark King Fahd University of Petroleum and Minerals Electrical Engineering Department Interdisciplinary Research Center for Sustainable Energy Systems Dhahran31261 Saudi Arabia Macquarie University School of Engineering NSW2019 Australia

This paper proposes a novel adaptive droop (ADC) control for voltage source converter (VSC)-based multiterminal medium-voltage direct current (MT-MVDC) distribution systems with heterogeneous energy storage system (ESS) participation. The proposed ADC scheme can achieve MVDC bus voltage management and power sharing between VSCs with overloading operation, solving the issue that all VSCs on the rectifier/inverter side reach their power limits concurrently after a large system disturbance. This is accomplished by factoring in the real-time power/voltage values at VSC stations along with an extended power component. This component is contingent on the actual operational state of the ESSs and inversely correlated with the power delivered/absorbed by the ESSs. Moreover, the controllers of the battery and ultracapacitor systems incorporate two additional control loops, a novel VSC excess power reduction (VEPR) loop as well as inner expanded excess power reduction (EEPR) loop, to obtain the power reference values. The two loops ensure all VSCs operate within individual power limits after the engagement of the ESS. The effectiveness of proposed ADC strategy is confirmed through real-time simulations in a hybrid AC/MT-MVDC system. © 1986-2012 IEEE.

关键词： Charge storage

来源：评论

学校读者我要写书评

暂无评论

Optimal first-order algorithms as a function of inequalities

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2024年第1期25卷 2557-2625页

作者： Chanwoo Park Ernest K. Ryu Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology Massachusetts Department of Mathematical Sciences Interdisciplinary Program in Artificial Intelligence Seoul National University Seoul Korea

In this work, we present a novel algorithm design methodology that finds the optimal algorithm as a function of inequalities. Specifically, we restrict convergence analyses of algorithms to use a prespecified subset of inequalities, rather than utilizing all true inequalities, and find the optimal algorithm subject to this restriction. This methodology allows us to design algorithms with certain desired characteristics. As concrete demonstrations of this methodology, we find new state-of-the-art accelerated first-order gradient methods using randomized coordinate updates and backtracking line searches.

关键词：

来源：评论

学校读者我要写书评

暂无评论

On the Impact of Knowledge Distillation for Model Interpretability 40

On the Impact of Knowledge Distillation for Model Interpreta...

引用

40th International Conference on Machine Learning, ICML 2023

作者： Han, Hyeongrok Siwon, Kim Choi, Hyun-Soo Yoon, Sungroh Department of Electrical and Computer Engineering Seoul National University Seoul Korea Republic of Department of Computer Science and Engineering Seoul National University of Science and Technology Seoul Korea Republic of ZIOVISION Inc. Chuncheon Korea Republic of Interdisciplinary Program in Artificial Intelligence Seoul National University Seoul Korea Republic of

Several recent studies have elucidated why knowledge distillation (KD) improves model performance. However, few have researched the other advantages of KD in addition to its improving model performance. In this study, we have attempted to show that KD enhances the interpretability as well as the accuracy of models. We measured the number of concept detectors identified in network dissection for a quantitative comparison of model interpretability. We attributed the improvement in interpretability to the class-similarity information transferred from the teacher to student models. First, we confirmed the transfer of class-similarity information from the teacher to student model via logit distillation. Then, we analyzed how class-similarity information affects model interpretability in terms of its presence or absence and degree of similarity information. We conducted various quantitative and qualitative experiments and examined the results on different datasets, different KD methods, and according to different measures of interpretability. Our research showed that KD models by large models could be used more reliably in various fields. The code is available at https://***/Rok07/KD_***. © 2023 Proceedings of Machine Learning Research. All rights reserved.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Neural Machine Translation Approach for Low-resource Languages using Long Short-term Memory Model

Neural Machine Translation Approach for Low-resource Languag...

引用

2023 International Conference on computer science, Information Technology and Engineering, ICCoSITE 2023

作者： Heryadi, Yaya Wijanarko, Bambang Dwi Murad, Dina Fitria Tho, Cuk Hashimoto, Kiyota Bina Nusantara University BINUS Graduate Program - Computer Science Computer Science Department Jakarta11480 Indonesia BINUS Online Learning Computer Science Department Bina Nusantara University Jakarta11480 Indonesia BINUS Online Learning Information Systems Department Bina Nusantara University Jakarta11480 Indonesia Prince of Songkla University Interdisciplinary Graduate School or Earth System Science and Andaman Natural Disaster Management Phuket83120 Thailand

ISBN: (纸本)9798350320954

Communication is a form of translation that human learns naturally since early childhood. Translating a language to another language has become instrumental when peoples interact with other people who speak a different language. One example of high-impact language translation is translating health protocols from the World Health Organization to many national languages to prevent the spread of the Covid-19 virus. However, the language translation is not an easy computation task when there is a language-resource gap. This paper presents empirical results on the performance of the Long Short-term Memory model as a machine translation involving Bahasa Indonesia and the Sundanese languages. Performance of the Sundanese language to Indonesian language translation is 0.93 average training and validation accuracy and 0.91 average testings BLEU score. Whilst, Indonesian language to Sundanese language translation is 0.92 average training and validation accuracy and 0.88 average testings BLEU score. Both models achieve 0.39 average training loss and 0.38 average validation loss. © 2023 IEEE.

关键词： Viruses

来源：评论

学校读者我要写书评

暂无评论

Recurrent Neural Network-based Models as Bahasa Indonesia-Sundanese Language Neural Machine Translator

Recurrent Neural Network-based Models as Bahasa Indonesia-Su...

引用

2023 International Conference on computer science, Information Technology and Engineering, ICCoSITE 2023

作者： Wijanarko, Bambang Dwi Heryadi, Yaya Murad, Dina Fitria Tho, Cuk Hashimoto, Kiyota BINUS Online Learning Computer Science Department Bina Nusantara University Jakarta11480 Indonesia BINUS Graduate Program - Doctor of Computer Science Computer Science Department Bina Nusantara University Jakarta11480 Indonesia BINUS Online Learning Information Systems Department Bina Nusantara University Jakarta11480 Indonesia Prince of Songkla University Interdisciplinary Graduate School or Earth System Science and Andaman Natural Disaster Management Phuket83120 Thailand

ISBN: (纸本)9798350320954

Sequential data are available in many domains including weather observation data, stock price, gene sequence, and many others. Sequential data processing is very instrumental as a solution to a variety of downstream tasks in Natural Language Processing such as language translation. Translating a language to another language has become instrumental when peoples interact with other people who speak a different language. However, language translation is not an easy computation task when there is a language-resource gap. This paper presents some empirical results from the exploration of the Recurrent Neural Network (RNN), the Gated Recurrent Unit (GRU), the Long Short-term Memory (LSTM), and the Bidirectional Long Short-term Memory (Bi-LSTM) models as machine translation models from Bahasa Indonesia to the Sundanese language. These empirical results show that the Bi-LSTM achieved the highest performance (0.95 average training accuracy and 0.92 average testings BLEU score) compared to the RNN (0.93 average training accuracy and 0.91 average testings BLEU score), the GRU (0.88 average training accuracy and 0.89 average testings BLEU score), and the LSTM (0.92 average training accuracy and 0.91 average testings BLEU score) models. These results validate some previously reported studies that claim the Bi-LSTM model potentially outperforms the other recurrent neural network-based models when it is used to process sequence datasets. © 2023 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Revalidating the Encoder-Decoder Depths and Activation Function to Find Optimum Vanilla Transformer Model

Revalidating the Encoder-Decoder Depths and Activation Funct...

引用

2023 International Conference on computer science, Information Technology and Engineering, ICCoSITE 2023

作者： Heryadi, Yaya Wijanarko, Bambang Dwi Fitria Murad, Dina Tho, Cuk Hashimoto, Kiyota Bina Nusantara University Graduate Program- Computer Science Computer Science Department BINUS Jakarta11480 Indonesia Bina Nusantara University Online Learning Computer Science Department BINUS Jakarta11480 Indonesia Bina Nusantara University BINUS Online Learning Information Systems Department Jakarta11480 Indonesia Prince of Songkla University Interdisciplinary Graduate School or Earth System Science and Andaman Natural Disaster Management Phuket83120 Thailand

ISBN: (纸本)9798350320954

The transformer model has become a state-of-the-art model in Natural Language Processing. The initial transformer model, known as the vanilla transformer model, is designed to improve some prominent models in sequence modeling and transduction problems such as language modeling and machine translation. The initial transformer model has 6 stacks of identical encoder-decoder layers with an attention mechanism whose aim is to push limitations of common recurrent language models and encoder-decoder architectures. Its outstanding performance has inspired many researchers to extend the architecture to improve its performance and computation efficiency. Despite many extensions to the vanilla transformer, there is no clear explanation of the encoder-decoder set out depth in the vanilla transformer model. This paper presents exploration results on the effect of combination encoder-decoder layer depth and activation function in the feed-forward layer of the vanilla transformer model on its performance. The model is tested to address a downstream task: text translation from Bahasa Indonesia to the Sundanese language. Although the value difference is not significantly large, the empirical results show that the combination of depth = 2 with Sigmoid, Tanh, and ReLU activation function;and d = 6 with ReLU activation shows the highest average training accuracy. Interestingly, d = 6 and ReLU show the lowest average training and validation loss. However, statistically, there is no significant difference between depth and activation functions. © 2023 IEEE.

关键词： Neural machine translation

来源：评论

学校读者我要写书评

暂无评论

Indonesian-Sundanese Language Machine Translation using Bidirectional Long Short-term Memory Model

Indonesian-Sundanese Language Machine Translation using Bidi...

引用

2023 International Conference on computer science, Information Technology and Engineering, ICCoSITE 2023

作者： Heryadi, Yaya Wijanarko, Bambang Dwi Fitria Murad, Dina Tho, Cuk Hashimoto, Kiyota BINUS Graduate Program - Doctor of Computer Science Computer Science Department Bina Nusantara University Jakarta11480 Indonesia BINUS Online Learning Bina Nusantara University Computer Science Department Jakarta11480 Indonesia BINUS Online Learning Information Systems Department Bina Nusantara University Jakarta11480 Indonesia Prince of Songkla University Interdisciplinary Graduate School or Earth System Science and Andaman Natural Disaster Management Phuket83120 Thailand

ISBN: (纸本)9798350320954

Translating a language to another language has become instrumental when peoples interact with other people who speak a different language. However, the language translation is not an easy computation task when there is a language-resource gap. This paper presents empirical results on the performance of two models: the Long Short-term Memory and the Bidirectional Long Short-term Memory models as machine language translation models involving Bahasa Indonesia and the Sundanese language. The empiric results showed that the Bidirectional Long Short-term Memory model achieves higher performance as a language translator from the Sundanese language to Bahasa Indonesia and vice versa (0.95 and 0.95 average training accuracy respectively;and 0.90 and 0.89 average testing BLEU scores respectively) than the Long Short-term Memory model as a language translator from the Sundanese language to Bahasa Indonesia and vice versa (0.93 and 0.92 average training accuracy respectively;and 0.91 and 0.88 average testing BLEU scores). These results validate some previously reported studies that claim the Bidirectional Long Short-term Memory model potentially outperform the Long Short-term Memory model when it is used to process a sequence dataset. © 2023 IEEE.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Photonic Crystals with Split Ring Unit Cells for Subwavelength Light Confinement

Photonic Crystals with Split Ring Unit Cells for Subwaveleng...

引用

CLEO: science and Innovations, S and I 2022

作者： Arnold, Kellen P. Halimi, Sami I. Allen, Joshua A. Hu, Shuren Weiss, Sharon M. Interdisciplinary Materials Science Program Vanderbilt University NashvilleTN37235 United States Department of Electrical Engineering and Computer Science Vanderbilt University NashvilleTN37235 United States

We report a split ring photonic crystal that demonstrates an order of magnitude larger peak energy density compared to traditional photonic crystals. The split ring offers highly focused optical energy in an accessibl...

ISBN: (纸本)9781557528209

关键词：

来源：评论

学校读者我要写书评

暂无评论

Reducing idleness in financial cloud services via multi-objective evolutionary reinforcement learning based load balancer

引用

science China(Information sciences) 2024年第2期67卷 20-40页

作者： Peng YANG Laoming ZHANG Haifeng LIU Guiying LI Department of Statistics and Data Science Southern University of Science and Technology Guangdong Provincial Key Laboratory of Brain-Inspired Intelligent Computation Department of Computer Science and EngineeringSouthern University of Science and Technology Academy for Advanced Interdisciplinary Studies Southern University of Science and Technology Guangdong OPPO Mobile Telecommunications Corp. Ltd Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology

In recent years, various companies have started to shift their data services from traditional data centers to the cloud. One of the major motivations is to save on operational costs with the aid of cloud elasticity. This paper discusses an emerging need from financial services to reduce the incidence of idle servers retaining very few user connections, without disconnecting them from the server side. This paper considers this need as a bi-objective online load balancing problem. A neural network based scalable policy is designed to route user requests to varied numbers of servers for the required elasticity. An evolutionary multi-objective training framework is proposed to optimize the weights of the policy. Not only is the new objective of idleness reduced by over 130% more than traditional industrial solutions, but the original load balancing objective itself is also slightly improved. Extensive simulations with both synthetic and real-world data help reveal the detailed applicability of the proposed method to the emergent problem of reducing idleness in financial services.

关键词： evolutionary reinforcement learning evolutionary multi-objective optimization load balance cloud computing

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：