检索结果-内蒙古大学图书馆

Cumulative adaptation for BLSTM acoustic models

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Kitza, Markus Golik, Pavel Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

This paper addresses the robust speech recognition problem as an adaptation task. Specifically, we investigate the cumulative application of adaptation methods. A bidirectional Long Short-Term Memory (BLSTM) based neural network, capable of learning temporal relationships and translation invariant representations, is used for robust acoustic modeling. Further, i-vectors were used as an input to the neural network to perform instantaneous speaker and environment adaptation, providing 8% relative improvement in word error rate on the NIST Hub5 2000 evaluation testset. By enhancing the first-pass i-vector based adaptation with a second-pass adaptation using speaker and environment dependent transformations within the network, a further relative improvement of 5% in word error rate was achieved. We have reevaluated the features used to estimate i-vectors and their normalization to achieve the best performance in a modern large scale automatic speech recognition system. Copyright © 2019, The Authors. All rights reserved.

关键词： Speech recognition

The rwth asr system for ted-lium release 2: improving hybrid hmm with specaugment

学校读者我要写书评

暂无评论

arXiv 2020年

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

We present a complete training pipeline to build a state-of-the-art hybrid HMM-based ASR system on the 2nd release of the TED-LIUM corpus. Data augmentation using SpecAugment is successfully applied to improve performance on top of our best SAT model using i-vectors. By investigating the effect of different maskings, we achieve improvements from SpecAugment on hybrid HMM models without increasing model size and training time. A subsequent sMBR training is applied to fine-tune the final acoustic model, and both LSTM and Transformer language models are trained and evaluated. Our best system achieves a 5.6% WER on the test set, which outperforms the previous state-of-the-art by 27% relative. Copyright © 2020, The Authors. All rights reserved.

关键词： Speech recognition

Improved training of end-to-end attention models for speech recognition

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Zeyer, Albert Irie, Kazuki Schlüter, Ralf Ney, Hermann Computer Science Department Rwth Aachen University Human Language Technology and Pattern Recognition Aachen52062 Germany AppTek United States Nnaisense Switzerland

Sequence-to-sequence attention-based models on subword units allow simple open-vocabulary end-to-end speech recognition. In this work, we show that such models can achieve competitive results on the Switchboard 300h and LibriSpeech 1000h tasks. In particular, we report the state-of-the-art word error rates (WER) of 3.54% on the dev-clean and 3.82% on the testclean evaluation subsets of LibriSpeech. We introduce a new pretraining scheme by starting with a high time reduction factor and lowering it during training, which is crucial both for convergence and final performance. In some experiments, we also use an auxiliary CTC loss function to help the convergence. In addition, we train long short-term memory (LSTM) language models on subword units. By shallow fusion, we report up to 27% relative improvements in WER over the attention baseline without a language model. Copyright © 2018, The Authors. All rights reserved.

关键词： Long short-term memory

Robust Beam Search for Encoder-Decoder Attention Based Speech recognition without Length Bias

学校读者我要写书评

暂无评论

arXiv 2020年

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

As one popular modeling approach for end-to-end speech recognition, attention-based encoder-decoder models are known to suffer the length bias and corresponding beam problem. Different approaches have been applied in simple beam search to ease the problem, most of which are heuristic-based and require considerable tuning. We show that heuristics are not proper modeling refinement, which results in severe performance degradation with largely increased beam sizes. We propose a novel beam search derived from reinterpreting the sequence posterior with an explicit length modeling. By applying the reinterpreted probability together with beam pruning, the obtained final probability leads to a robust model modification, which allows reliable comparison among output sequences of different lengths. Experimental verification on the LibriSpeech corpus shows that the proposed approach solves the length bias problem without heuristics or additional tuning effort. It provides robust decision making and consistently good performance under both small and very large beam sizes. Compared with the best results of the heuristic baseline, the proposed approach achieves the same WER on the 'clean' sets and 4% relative improvement on the 'other' sets. We also show that it is more efficient with the additional derived early stopping criterion. Copyright © 2020, The Authors. All rights reserved.

关键词： Speech recognition

A systematic comparison of grapheme-based vs. phoneme-based label units for encoder-decoder-attention models

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zeineldeen, Mohammad Zeyer, Albert Zhou, Wei Ng, Thomas Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University 52062 Aachen Germany AppTek GmbH Aachen52062 Germany

Following the rationale of end-to-end modeling, CTC, RNN-T or encoder-decoder-attention models for automatic speech recognition (ASR) use graphemes or grapheme-based subword units based on e.g. byte-pair encoding (BPE). The mapping from pronunciation to spelling is learned completely from data. In contrast to this, classical approaches to ASR employ secondary knowledge sources in the form of phoneme lists to define phonetic output labels and pronunciation lexica. In this work, we do a systematic comparison between grapheme- and phoneme-based output labels for an encoder-decoder-attention ASR model. We investigate the use of single phonemes as well as BPE-based phoneme groups as output labels of our model. To preserve a simplified and efficient decoder design, we also extend the phoneme set by auxiliary units to be able to distinguish homophones. Experiments performed on the Switchboard 300h and LibriSpeech benchmarks show that phoneme-based modeling is competitive to grapheme-based encoder-decoder-attention modeling. Copyright © 2020, The Authors. All rights reserved.

关键词： Speech recognition

Comparing the benefit of synthetic training data for various automatic speech recognition architectures

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Rossenbach, Nick Zeineldeen, Mohammad Hilmes, Benedikt Schlüter, Ralf Ney, Hermann Human Language Technology and Pattern Recognition Computer Science Department RWTH Aachen University Aachen52074 Germany AppTek GmbH Aachen52062 Germany

Recent publications on automatic-speech-recognition (ASR) have a strong focus on attention encoder-decoder (AED) architectures which tend to suffer from over-fitting in low resource scenarios. One solution to tackle this issue is to generate synthetic data with a trained text-to-speech system (TTS) if additional text is available. This was successfully applied in many publications with AED systems, but only very limited in the context of other ASR architectures. We investigate the effect of varying pre-processing, the speaker embedding and input encoding of the TTS system w.r.t. the effectiveness of the synthesized data for AED-ASR training. Additionally, we also consider internal language model subtraction for the first time, resulting in up to 38% relative improvement. We compare the AED results to a state-of-the-art hybrid ASR system, a monophone based system using connectionist-temporal-classification (CTC) and a monotonic transducer based system. We show that for the later systems the addition of synthetic data has no relevant effect, but they still outperform the AED systems on LibriSpeech-100h. We achieve a final word-error-rate of 3.3%/10.0% with a hybrid system on the clean/noisy test-sets, surpassing any previous state-of-the-art systems on Librispeech-100h that do not include unlabeled audio data. Copyright © 2021, The Authors. All rights reserved.

关键词： Speech recognition

Confidence scores for acoustic model adaptation

学校读者我要写书评

暂无评论

Confidence scores for acoustic model adaptation

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

作者： Christian Gollan Michiel Bacchiani Human Language Technology and Pattern Recognition Computer Science Department 6 RWTH Aachen University Germany Google Inc. New York NY USA

This paper focuses on confidence scores for use in acoustic model adaptation. Frame-based confidence estimates are used in linear transform (CMLLR and MLLR) and MAP adaptation. We show that adaptation approaches with a limited number of free parameters such as linear transform-based approaches are robust in the face of frame labeling errors whereas adaptation approaches with a large number of free parameters such as MAP are sensitive to the quality of the supervision and hence benefit most from use of confidences. Different approaches for using confidence information in adaptation are investigated. This analysis shows that a thresholding approach is effective in that it improves the frame labeling accuracy with little detrimental effect on frame recall. Experimental results show an absolute WER reduction of 2.1% over a CMLLR adapted system on a video transcription task.

关键词： Adaptation model Maximum likelihood linear regression Parameter estimation Speech Robustness Labeling Multimedia systems humans pattern recognition Computer science

ROBUST KNOWLEDGE DISTILLATION FROM RNN-T MODELS WITH NOISY TRAINING LABELS USING FULL-SUM LOSS

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zeineldeen, Mohammad Audhkhasi, Kartik Baskar, Murali Karthick Ramabhadran, Bhuvana Human Language Technology and Pattern Recognition Computer Science Department Rwth Aachen University Aachen52074 Germany Google Llc New York United States

This work studies knowledge distillation (KD) and addresses its constraints for recurrent neural network transducer (RNNT) models. In hard distillation, a teacher model transcribes large amounts of unlabelled speech to train a student model. Soft distillation is another popular KD method that distills the output logits of the teacher model. Due to the nature of RNN-T alignments, applying soft distillation between RNNT architectures having different posterior distributions is challenging. In addition, bad teachers having high word-error-rate (WER) reduce the efficacy of KD. We investigate how to effectively distill knowledge from variable quality ASR teachers, which has not been studied before to the best of our knowledge. We show that a sequence-level KD, full-sum distillation, outperforms other distillation methods for RNN-T models, especially for bad teachers. We also propose a variant of full-sum distillation that distills the sequence discriminative knowledge of the teacher leading to further improvement in WER. We conduct experiments on public datasets namely SpeechStew and LibriSpeech, and on in-house production data. Copyright © 2023, The Authors. All rights reserved.

关键词： Recurrent neural networks

A comparative large scale study of MLP features for mandarin ASR

学校读者我要写书评

暂无评论

A comparative large scale study of MLP features for mandarin...

作者： Valente, Fabio Doss, Mathew Magimai Plahl, Christian Ravuri, Suman Wang, Wen IDIAP Research Institute CH-1920 Martigny Switzerland Human Language Technology and Pattern Recognition RWTH Aachen University Germany International Computer Science Institute 1947 Center Street Berkeley CA 94704 United States Speech Technology and Research Laboratory SRI International Menlo Park CA United States

MLP based front-ends have shown significant complementary properties to conventional spectral features. As part of the DARPA GALE program, different MLP features were developed for Mandarin ASR. In this paper, all the proposed frontends are compared in systematic manner and we extensively investigate the scalability of these features in terms of the amount of training data (from 100 hours to 1600 hours) and system complexity (maximum likelihood training, SAT, lattice level combination, and discriminative training). Results on 5 hours of evaluation data from the GALE project reveal that the MLP features consistently produce relative improvements in the range of 15% - 23% at the different steps of a multipass system when compared to the conventional short-term spectral based features like MFCC and PLP. The largest improvement is obtained using a hierarchical MLP approach. © 2010 ISCA.

关键词： Maximum likelihood