检索结果-内蒙古大学图书馆

Why does CTC result in peaky behavior?

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zeyer, Albert Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen Germany AppTek GmbH Aachen Germany

The peaky behavior of CTC models is well known experimentally. However, an understanding about why peaky behavior occurs is missing, and whether this is a good property. We provide a formal analysis of the peaky behavior and gradient descent convergence properties of the CTC loss and related training criteria. Our analysis provides a deep understanding why peaky behavior occurs and when it is suboptimal. On a simple example which should be trivial to learn for any model, we prove that a feed-forward neural network trained with CTC from uniform initialization converges towards peaky behavior with a 100% error rate. Our analysis further explains why CTC only works well together with the blank label. We further demonstrate that peaky behavior does not occur on other related losses including a label prior model, and that this improves convergence. © 2021, CC BY-SA.

关键词： Gradient methods

One decade of statistical machine translation: 1996-2005

学校读者我要写书评

暂无评论

One decade of statistical machine translation: 1996-2005

IEEE Workshop on Automatic Speech recognition and Understanding

作者： H. Ney Human Language Technology and Pattern Recognition Lehrstuhl fur Informatik VI-Computer Science Department RWTH Aachen University Aachen Germany

In the last decade, the statistical approach has found widespread use in machine translation both for written and spoken language and has had a major impact on the translation accuracy. The goal of this paper is to cover the state of the art in statistical machine translation. We would re-visit the underlying principles of the statistical approach to machine translation and summarize the progress that has been made over the last decade

关键词： Natural languages Context modeling Hidden Markov models Speech humans pattern recognition computer science Transducers Tides Vocabulary

THE RWTH ENGLISH LECTURE recognition SYSTEM

学校读者我要写书评

暂无评论

THE RWTH ENGLISH LECTURE RECOGNITION SYSTEM

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Simon Wiesler Kazuki Irie Zoltan Tuske Ralf Schluter Hermann Ney Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen Germany Ecole Centrale Paris Paris France

ISBN: (纸本)9781479928941

In this paper, we describe the RWTH speech recognition system for English lectures developed within the Translectures project. A difficulty in the development of an English lectures recognition system, is the high ratio of non-native speakers. We address this problem by using very effective deep bottleneck features trained on multilingual data. The acoustic model is trained on large amounts of data from different domains and with different dialects. Large improvements are obtained from unsupervised acoustic adaptation. Another challenge is the frequent use of technical terms and the wide range of topics. In our recognition system, slides, which are attached to most lectures, are used for improving lexical coverage and language model adaptation.

关键词： Lectures (teaching method) recognition system modelling languages speech recognition system English English language Acoustic models

Acoustic feature combination for robust speech recognition

学校读者我要写书评

暂无评论

Acoustic feature combination for robust speech recognition

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： A. Zolnay R. Schluter H. Ney Human Language Technology and Pattern Recognition Lehrstuhl für Informatik VI Computer Science Department RWTH Aachen University Aachen Germany

In this paper, we consider the use of multiple acoustic features of the speech signal for robust speech recognition. We investigate the combination of various auditory based (mel frequency cepstrum coefficients, perceptual linear prediction, etc.) and articulatory based (voicedness) features. Features are combined by linear discriminant analysis and log-linear model combination based techniques. We describe the two feature combination techniques and compare the experimental results. Experiments performed on the large-vocabulary task VerbMobil II (German conversational speech) show that the accuracy of automatic speech recognition systems can be improved by the combination of different acoustic features.

关键词： Robustness Speech recognition Mel frequency cepstral coefficient Linear discriminant analysis Cepstrum Feature extraction Automatic speech recognition Autocorrelation Cepstral analysis Concatenated codes

The RWTH Arabic-to-English spoken language translation system

学校读者我要写书评

暂无评论

The RWTH Arabic-to-English spoken language translation syste...

IEEE Workshop on Automatic Speech recognition and Understanding

作者： Oliver Bender Evgeny Matusov Stefan Hahn Sasa Hasan Shahram Khadivi Hermann Ney Human Language Technology and Pattern Recognition Lehrstuhl für Informatik 6-Computer Science Department RWTH Aachen University Aachen Germany

ISBN: (纸本)9781424413690;1424413699

We present the RWTH phrase-based statistical machine translation system designed for the translation of Arabic speech into English text. This system was used in the Global Autonomous language Exploitation (GALE) Go/No-Go Translation Evaluation 2007. Using a two-pass approach, we first generate n-best translation candidates and then rerank these candidates using additional models. We give a short review of the decoder as well as of the models used in both passes. We stress the difficulties of spoken language translation, i.e. how to combine the recognition and translation systems and how to compensate for missing punctuation. In addition, we cover our work on domain adaptation for the applied language models. We present translation results for the official GALE 2006 evaluation set and the GALE 2007 development set.

关键词： Natural languages Automatic speech recognition Surface-mount technology Vocabulary Decoding Speech analysis Broadcasting humans pattern recognition computer science

On the choice of modeling unit for sequence-to-sequence speech recognition

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Irie, Kazuki Prabhavalkar, Rohit Kannan, Anjuli Bruguier, Antoine Rybach, David Nguyen, Patrick Human Language Technology and Pattern Recognition Group Computer Science Department RWTH Aachen University AachenD-52056 Germany Google Mountain ViewCA94043 United States

Efficient Training of Neural Transducer for Speech recognition

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhou, Wei Michel, Wilfried Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

As one of the most popular sequence-to-sequence modeling approaches for speech recognition, the RNN-Transducer has achieved evolving performance with more and more sophisticated neural network models of growing size and increasing training epochs. While strong computation resources seem to be the prerequisite of training superior models, we try to overcome it by carefully designing a more efficient training pipeline. In this work, we propose an efficient 3-stage progressive training pipeline to build highly-performing neural transducer models from scratch with very limited computation resources in a reasonable short time period. The effectiveness of each stage is experimentally verified on both Librispeech and Switchboard corpora. The proposed pipeline is able to train transducer models approaching state-of-the-art performance with a single GPU in just 2-3 weeks. Our best conformer transducer achieves 4.1% WER on Librispeech test-other with only 35 epochs of training. Copyright © 2022, The Authors. All rights reserved.

关键词： Speech recognition

Analysis of deep clustering as preprocessing for automatic speech recognition of sparsely overlapping speech

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Menne, Tobias Sklyar, Ilya Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

Significant performance degradation of automatic speech recognition (ASR) systems is observed when the audio signal contains cross-talk. One of the recently proposed approaches to solve the problem of multi-speaker ASR is the deep clustering (DPCL) approach. Combining DPCL with a state-of-the-art hybrid acoustic model, we obtain a word error rate (WER) of 16.5 % on the commonly used wsj0-2mix dataset, which is the best performance reported thus far to the best of our knowledge. The wsj0-2mix dataset contains simulated cross-talk where the speech of multiple speakers overlaps for almost the entire utterance. In a more realistic ASR scenario the audio signal contains significant portions of single-speaker speech and only part of the signal contains speech of multiple competing speakers. This paper investigates obstacles of applying DPCL as a preprocessing method for ASR in such a scenario of sparsely overlapping speech. To this end we present a data simulation approach, closely related to the wsj0-2mix dataset, generating sparsely overlapping speech datasets of arbitrary overlap ratio. The analysis of applying DPCL to sparsely overlapping speech is an important interim step between the fully overlapping datasets like wsj0-2mix and more realistic ASR datasets, such as CHiME-5 or AMI. Copyright © 2019, The Authors. All rights reserved.

关键词： Speech

Phoneme based neural transducer for large vocabulary speech recognition

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zhou, Wei Berger, Simon Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

To join the advantages of classical and end-to-end approaches for speech recognition, we present a simple, novel and competitive approach for phoneme-based neural transducer modeling. Different alignment label topologies are compared and word-end-based phoneme label augmentation is proposed to improve performance. Utilizing the local dependency of phonemes, we adopt a simplified neural network structure and a straightforward integration with the external word-level language model to preserve the consistency of seq-to-seq modeling. We also present a simple, stable and efficient training procedure using frame-wise cross-entropy loss. A phonetic context size of one is shown to be sufficient for the best performance. A simplified scheduled sampling approach is applied for further improvement and different decoding approaches are briefly compared. The overall performance of our best model is comparable to state-of-the-art (SOTA) results for the TED-LIUM Release 2 and Switchboard corpora. Copyright © 2020, The Authors. All rights reserved.

关键词： Transducers