检索结果-内蒙古大学图书馆

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Albert Zeyer Patrick Doetsch Paul Voigtlaender Ralf Schluter Hermann Ney Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University 52062 Germany

ISBN: (纸本)9781509041183

Recent experiments show that deep bidirectional long short-term memory (BLSTM) recurrent neural network acoustic models outperform feedforward neural networks for automatic speech recognition (ASR). However, their training requires a lot of tuning and experience. In this work, we provide a comprehensive overview over various BLSTM training aspects and their interplay within ASR, which has been missing so far in the literature. We investigate on different variants of optimization methods, batching, truncated backpropagation, and regularization techniques such as dropout, and we study the effect of size and depth, training models of up to 10 layers. This includes a comparison of computation times vs. recognition performance. Furthermore, we introduce a pretraining scheme for LSTMs with layer-wise construction of the network showing good improvements especially for deep networks. The experimental analysis mainly was performed on the Quaero task, with additional results on Switchboard. The best BLSTM model gave a relative improvement in word error rate of over 15% compared to our best feed-forward baseline on our Quaero 50h task. All experiments were done using RETURNN and RASR, RWTH's extensible training framework for universal recurrent neural networks and ASR toolkit. The training configuration files are publicly available.

关键词： acoustic modeling LSTM RNN acoustic modeling Speech recognition recurrent neural nets aquifer storage and recovery Functional training feedforward neural nets Bidirectional

来源：评论

学校读者我要写书评

暂无评论

Local system voting feature for machine translation system combination

arXiv

引用

arXiv 2017年

作者： Freitag, Markus Peter, Jan-Thorsten Peitz, Stephan Feng, Minwei Ney, Hermann Human Language Technology Pattern Recognition Group Computer Science Department RWTH Aachen University AachenD-52056 Germany

In this paper, we enhance the traditional confusion network system combination approach with an additional model trained by a neural network. This work is motivated by the fact that the commonly used binary system voting models only assign each input system a global weight which is responsible for the global impact of each input system on all translations. This prevents individual systems with low system weights from having influence on the system combination output, although in some situations this could be helpful. Further, words which have only been seen by one or few systems rarely have a chance of being present in the combined output. We train a local system voting model by a neural network which is based on the words themselves and the combinatorial occurrences of the different system outputs. This gives system combination the option to prefer other systems at different word positions even for the same sentence. Copyright © 2017, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

CharacTER: Translation Edit Rate on Character Level 1

CharacTER: Translation Edit Rate on Character Level

引用

1st Conference on Machine Translation, WMT 2016, held at the 54th Annual Meeting of the Association for Computational Linguistics, ACL 2016

作者： Wang, Weiyue Peter, Jan-Thorsten Rosendahl, Hendrik Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52056 Germany

ISBN: (纸本)9781945626104

Recently, the capability of character-level evaluation measures for machine translation output has been confirmed by several metrics. This work proposes translation edit rate on character level (CharacTER), which calculates the character level edit distance while performing the shift edit on word level. The novel metric shows high system-level correlation with human rankings, especially for morphologically rich languages. It outperforms the strong CHRF by up to 7% correlation on different metric tasks. In addition, we apply the hypothesis sentence length for normalizing the edit distance in CharacTER, which also provides significant improvements compared to using the reference sentence length. © 2016 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Global investments in pandemic preparedness and COVID-19: development assistance and domestic spending on health between 1990 and 2026

引用

The Lancet Global Health 2023年第3期11卷 e385-e413页

作者： Micah, Angela E. Bhangdia, Kayleigh Cogswell, Ian E. Lasher, Dylan Lidral-Porter, Brendan Maddison, Emilie R. Nguyen, Trang Nhu Ngoc Patel, Nishali Pedroza, Paola Solorio, Juan Stutzman, Hayley Tsakalos, Golsum Wang, Yifeng Warriner, Wesley Zhao, Yingxi Zlavog, Bianca S. Abbafati, Cristiana Abbas, Jaffar Abbasi-Kangevari, Mohsen Abbasi-Kangevari, Zeinab Abdelmasseh, Michael Abdulah, Deldar Morad Abedi, Aidin Abegaz, Kedir Hussein Abhilash, E.S. Aboagye, Richard Gyan Abolhassani, Hassan Abrigo, Michael R.M. Abubaker Ali, Hiwa Abu-Gharbieh, Eman Adem, Mohammed Hussien Afzal, Muhammad Sohail Ahmadi, Ali Ahmed, Haroon Ahmed Rashid, Tarik Aji, Budi Akbarialiabad, Hossein Akelew, Yibeltal Al Hamad, Hanadi Alam, Khurshid Alanezi, Fahad Mashhour Alanzi, Turki M. Al-Hanawi, Mohammed Khaled Alhassan, Robert Kaba Aljunid, Syed Mohamed Almustanyir, Sami Al-Raddadi, Rajaa M. Alvis-Guzman, Nelson Alvis-Zakzuk, Nelson J. Amare, Azmeraw T. Ameyaw, Edward Kwabena Amini-Rarani, Mostafa Amu, Hubert Ancuceanu, Robert Andrei, Tudorel Anwar, Sumadi Lukman Appiah, Francis Aqeel, Muhammad Arabloo, Jalal Arab-Zozani, Morteza Aravkin, Aleksandr Y. Aremu, Olatunde Aruleba, Raphael Taiwo Athari, Seyyed Shamsadin Avila-Burgos, Leticia Ayanore, Martin Amogre Azari, Samad Baig, Atif Amin Bantie, Abere Tilahun Barrow, Amadou Baskaran, Pritish Basu, Sanjay Batiha, Abdul-Monim Mohammad Baune, Bernhard T. Berezvai, Zombor Bhardwaj, Nikha Bhardwaj, Pankaj Bhaskar, Sonu Boachie, Micheal Kofi Bodolica, Virginia Botelho, João Silva Botelho Braithwaite, Dejana Breitborde, Nicholas J.K. Busse, Reinhard Cahuana-Hurtado, Lucero Catalá-López, Ferrán Chansa, Collins Charan, Jaykaran Chattu, Vijay Kumar Chen, Simiao Chukwu, Isaac Sunday Dadras, Omid Dandona, Lalit Dandona, Rakhi Dargahi, Abdollah Debela, Sisay Abebe Denova-Gutiérrez, Edgar Desye, Belay Dharmaratne, Samath Dhamminda Diao, Nancy Doan, Linh Phuong Dodangeh, Milad dos Santos, Wendel Mombaque Doshmangir, Leila Dube, John Eini, Ebrahim El Sayed Zaki, Maysaa El Tantawi, Maha Enyew, Daniel Berhanie Institute for Health Metrics and Evaluation University of Washington Seattle WA United States Department of Applied Mathematics University of Washington Seattle WA United States Department of Health Metrics Sciences School of Medicine University of Washington Seattle WA United States Nuffield Department of Medicine University of Oxford Oxford United Kingdom Department of Juridical and Economic Studies La Sapienza University Rome Italy Antai College of Economics Shanghai Jiao Tong University Shanghai China Non-communicable Diseases Research Center Tehran University of Medical Sciences Tehran Iran Research Center for Immunodeficiencies Tehran University of Medical Sciences Tehran Iran Multiple Sclerosis Research Center Tehran University of Medical Sciences Tehran Iran Children’s Medical Center Tehran University of Medical Sciences Tehran Iran Department of Epidemiology and Biostatistics Tehran University of Medical Sciences Tehran Iran Tehran Heart Center Tehran University of Medical Sciences Tehran Iran Faculty of Medicine Tehran University of Medical Sciences Tehran Iran National Institute for Health Research Tehran University of Medical Sciences Tehran Iran Department of Pharmacology Tehran University of Medical Sciences Tehran Iran Department of Cardiology Tehran University of Medical Sciences Tehran Iran Sina Trauma and Surgery Research Center Tehran University of Medical Sciences Tehran Iran Digestive Diseases Research Institute Tehran University of Medical Sciences Tehran Iran Department of Neurology Tehran University of Medical Sciences Tehran Iran Social Determinants of Health Research Center Shahid Beheshti University of Medical Sciences Tehran Iran Department of Epidemiology Shahid Beheshti University of Medical Sciences Tehran Iran Department of Health Economics and Statistics Shahid Beheshti University of Medical Sciences Tehran Iran Department of Neurosurgery Shahid Beheshti University of Medical Sciences Tehran Iran Ophthalmic R

Background: The COVID-19 pandemic highlighted gaps in health surveillance systems, disease prevention, and treatment globally. Among the many factors that might have led to these gaps is the issue of the financing of national health systems, especially in low-income and middle-income countries (LMICs), as well as a robust global system for pandemic preparedness. We aimed to provide a comparative assessment of global health spending at the onset of the pandemic;characterise the amount of development assistance for pandemic preparedness and response disbursed in the first 2 years of the COVID-19 pandemic;and examine expectations for future health spending and put into context the expected need for investment in pandemic preparedness. Methods: In this analysis of global health spending between 1990 and 2021, and prediction from 2021 to 2026, we estimated four sources of health spending: development assistance for health (DAH), government spending, out-of-pocket spending, and prepaid private spending across 204 countries and territories. We used the Organisation for Economic Co-operation and Development (OECD)'s Creditor Reporting System (CRS) and the WHO Global Health Expenditure Database (GHED) to estimate spending. We estimated development assistance for general health, COVID-19 response, and pandemic preparedness and response using a keyword search. Health spending estimates were combined with estimates of resources needed for pandemic prevention and preparedness to analyse future health spending patterns, relative to need. Findings: In 2019, at the onset of the COVID-19 pandemic, US$9·2 trillion (95% uncertainty interval [UI] 9·1–9·3) was spent on health worldwide. We found great disparities in the amount of resources devoted to health, with high-income countries spending $7·3 trillion (95% UI 7·2–7·4) in 2019;293·7 times the $24·8 billion (95% UI 24·3–25·3) spent by low-income countries in 2019. That same year, $43·1 billion in development assistance was provided

关键词：

来源：评论

学校读者我要写书评

暂无评论

The RWTH Aachen University English-Romanian Machine Translation System for WMT 2016 1

The RWTH Aachen University English-Romanian Machine Translat...

引用

1st Conference on Machine Translation, WMT 2016, held at the 54th Annual Meeting of the Association for Computational Linguistics, ACL 2016

作者： Peter, Jan-Thorsten Alkhouli, Tamer Guta, Andreas Ney, Hermann Human Language Technology and Pattern Recognition Group Computer Science Department RWTH Aachen University AachenD-52056 Germany

ISBN: (纸本)9781945626104

This paper describes the statistical machine translation system developed at RWTH Aachen University for the English→Romanian translation task of the ACL 2016 First Conference on Machine Translation (WMT 2016). We combined three different state-of the- art systems in a system combination: A phrase-based system, a hierarchical phrase-based system and an attentionbased neural machine translation system. The phrase-based and the hierarchical phrase-based systems make use of a language model trained on all available data, a language model trained on the bilingual data and a word class language model. In addition, we utilized a recurrent neural network language model and a bidirectional recurrent neural network translation model for reranking the output of both systems. The attention-based neural machine translation system was trained using all bilingual data together with the backtranslated data from the News Crawl 2015 corpora. © 2016 Association for Computational Linguistics.

关键词： Neural machine translation

来源：评论

学校读者我要写书评

暂无评论

Bridging the Native language and language Variety Identification Tasks

引用

Procedia computer science 2017年 112卷 1554-1561页

作者： Marc Franco-Salvador Greg Kondrak Paolo Rosso Symanto Research 90425 Nuremberg Germany Pattern Recognition and Human Language Technology (PRHLT) Research Center Universitat Politècnica de València 46022 Valencia Spain Department of Computing Science University of Alberta Edmonton AB T6G 2E8 Canada

The objective of Native language Identification is to determine the native language of the author of a text that he or she wrote in another language. By contrast, language Variety Identification aims at classifying texts representing different varieties of a single language. We postulate that both tasks may be reduced to a single objective, which is to identify the language variety of the text. We design a general approach that combines string kernels and word embeddings, which capture different characteristics of texts. The results of our experiments show that the approach achieves excellent results on both tasks, without any task-specific adaptations.

关键词： Native language Identification language Variety Identification String Kernels Word Embeddings Classifier Combination

来源：评论

学校读者我要写书评

暂无评论

On the Benefits of Convolutional Neural Network Combinations in Offline Handwriting recognition

On the Benefits of Convolutional Neural Network Combinations...

引用

International Workshop on Frontiers in Handwriting recognition

作者： Dewi Suryani Patrick Doetsch Hermann Ney King Mongkut's University of Technology North Bangkok The Sirindhorn International Thai-German Graduate School of Engineering Bangkok Thailand Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen Germany

ISBN: (纸本)9781509009824

In this paper, we elaborate the advantages of combining two neural network methodologies, convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent neural networks, with the framework of hybrid hidden Markov models (HMM) for recognizing offline handwriting text. CNNs employ shift-invariant filters to generate discriminative features within neural networks. We show that CNNs are powerful tools to extract general purpose features that even work well for unknown classes. We evaluate our system on a Chinese handwritten text database and provide a GPU-based implementation that can be used to reproduce the experiments. All experiments were conducted with RWTH OCR, an open-source system developed at our institute.

关键词： Hidden Markov models Neurons Training Handwriting recognition Biological neural networks computer architecture Logic gates

来源：评论

学校读者我要写书评

暂无评论

Local system voting feature for machine translation system combination 10

Local system voting feature for machine translation system c...

引用

10th Workshop on Statistical Machine Translation, WMT 2015 at the 2015 Conference on Empirical Methods in Natural language Processing, EMNLP 2015

作者： Freitag, Markus Peter, Jan-Thorsten Peitz, Stephan Feng, Minwei Ney, Hermann Human Language Technology and Pattern Recognition Group Computer Science Department Rwth Aachen University AachenD-52056 Germany

ISBN: (纸本)9781941643327

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

The rwth aachen german-english machine translation system for wmt 2015 10

The rwth aachen german-english machine translation system fo...

引用

10th Workshop on Statistical Machine Translation, WMT 2015 at the 2015 Conference on Empirical Methods in Natural language Processing, EMNLP 2015

作者： Peter, Jan-Thorsten Toutounchi, Farzad Wuebker, Joern Ney, Hermann Human Language Technology and Pattern Recognition Group Computer Science Department Rwth Aachen University AachenD-52056 Germany

ISBN: (纸本)9781941643327

This paper describes the statistical machine translation system developed at RWTH Aachen University for the German!English translation task of the EMNLP 2015 Tenth Workshop on Statistical Machine Translation (WMT 2015). A phrase-based machine translation system was applied and augmented with hierarchical phrase reordering and word class language models. Further, we ran discriminative maximum expected BLEU training for our system. In addition, we utilized multiple feed-forward neural network language and translation models and a recurrent neural network language model for reranking. © EMNLP 2015. All rights reserved.

关键词： Feedforward neural networks

来源：评论

学校读者我要写书评

暂无评论

Investigation of mixture splitting concept for training linear bottlenecks of deep neural network acoustic models 40

Investigation of mixture splitting concept for training line...

引用

40th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015

作者： Tahir, Muhammad Ali Wiesler, Simon Schluter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Germany Spoken Language Processing Group LIMSI CNRS Paris France

ISBN: (纸本)9781467369978

A Gaussian or log-linear mixture model trained by maximum likelihood may be trained further using discriminative training. It is desirable that the mixture splitting is also done during the discriminative training, to achieve better mixture density distribution. In previous work such a discriminative splitting approach was presented. Similarly, the resolution of a deep neural network may also be increased by splitting. In this paper, discriminative splitting is applied as a way of initializing a linear bottleneck between two layers of a DNN. Experiments for a single hidden layer and six hidden layer cases show the potential of this approach as an alternative method of pre-training for linear bottlenecks for MLP hidden layers. © 2015 IEEE.

关键词： deep neural network linear bottleneck mixture splitting

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：