English
Related papers

Related papers: RNN-Transducer-based Losses for Speech Recognition…

200 papers

Confidence estimate is an often requested feature in applications such as medical transcription where errors can impact patient care and the confidence estimate could be used to alert medical professionals to verify potential errors in…

Computation and Language · Computer Science 2021-10-29 Mingqiu Wang , Hagen Soltau , Laurent El Shafey , Izhak Shafran

We investigate a set of techniques for RNN Transducers (RNN-Ts) that were instrumental in lowering the word error rate on three different tasks (Switchboard 300 hours, conversational Spanish 780 hours and conversational Italian 900 hours).…

Computation and Language · Computer Science 2021-03-19 George Saon , Zoltan Tueske , Daniel Bolanos , Brian Kingsbury

This paper introduces a novel Token-and-Duration Transducer (TDT) architecture for sequence-to-sequence tasks. TDT extends conventional RNN-Transducer architectures by jointly predicting both a token and its duration, i.e. the number of…

Audio and Speech Processing · Electrical Eng. & Systems 2023-05-31 Hainan Xu , Fei Jia , Somshubra Majumdar , He Huang , Shinji Watanabe , Boris Ginsburg

The two most popular loss functions for streaming end-to-end automatic speech recognition (ASR) are RNN-Transducer (RNN-T) and connectionist temporal classification (CTC). Between these two loss types we can classify the monotonic RNN-T…

Audio and Speech Processing · Electrical Eng. & Systems 2022-10-25 Niko Moritz , Frank Seide , Duc Le , Jay Mahadeokar , Christian Fuegen

Transducer-based models, such as RNN-Transducer and transformer-transducer, have achieved great success in speech recognition. A typical transducer model decodes the output sequence conditioned on the current acoustic state and previously…

Audio and Speech Processing · Electrical Eng. & Systems 2021-04-08 Zhengkun Tian , Jiangyan Yi , Ye Bai , Jianhua Tao , Shuai Zhang , Zhengqi Wen

RNN-T models are widely used in ASR, which rely on the RNN-T loss to achieve length alignment between input audio and target sequence. However, the implementation complexity and the alignment-based optimization target of RNN-T loss lead to…

Sound · Computer Science 2024-11-28 Tian-Hao Zhang , Dinghao Zhou , Guiping Zhong , Jiaming Zhou , Baoxiang Li

The Recurrent Neural Network-Transducer (RNN-T) is widely adopted in end-to-end (E2E) automatic speech recognition (ASR) tasks but depends heavily on large-scale, high-quality annotated data, which are often costly and difficult to obtain.…

Computation and Language · Computer Science 2025-11-07 Dongji Gao , Chenda Liao , Changliang Liu , Matthew Wiesner , Leibny Paola Garcia , Daniel Povey , Sanjeev Khudanpur , Jian Wu

Knowledge Distillation is an effective method of transferring knowledge from a large model to a smaller model. Distillation can be viewed as a type of model compression, and has played an important role for on-device ASR applications. In…

Audio and Speech Processing · Electrical Eng. & Systems 2020-11-13 Sankaran Panchapagesan , Daniel S. Park , Chung-Cheng Chiu , Yuan Shangguan , Qiao Liang , Alexander Gruenstein

Transformers (Vaswani et al., 2017) have brought a remarkable improvement in the performance of neural machine translation (NMT) systems but they could be surprisingly vulnerable to noise. In this work, we try to investigate how noise…

Computation and Language · Computer Science 2021-09-13 Peyman Passban , Puneeth S. M. Saladi , Qun Liu

The transcriptions used to train an Automatic Speech Recognition (ASR) system may contain errors. Usually, either a quality control stage discards transcriptions with too many errors, or the noisy transcriptions are used as is. We introduce…

Computation and Language · Computer Science 2019-10-17 Adrien Dufraux , Emmanuel Vincent , Awni Hannun , Armelle Brun , Matthijs Douze

A significant performance reduction is often observed in speech recognition when the rate of speech (ROS) is too low or too high. Most of present approaches to addressing the ROS variation focus on the change of speech signals in dynamic…

Computation and Language · Computer Science 2015-06-03 Xiangyu Zeng , Shi Yin , Dong Wang

The loss function is crucial to machine learning, especially in supervised learning frameworks. It is a fundamental component that controls the behavior and general efficacy of learning algorithms. However, despite their widespread use,…

Machine Learning · Computer Science 2026-02-09 Soumi Mahato , Lineesh M. C

Recurrent Neural Network Transducer (RNN-T), like most end-to-end speech recognition model architectures, has an implicit neural network language model (NNLM) and cannot easily leverage unpaired text data during training. Previous work has…

Computation and Language · Computer Science 2020-10-28 Suyoun Kim , Yuan Shangguan , Jay Mahadeokar , Antoine Bruguier , Christian Fuegen , Michael L. Seltzer , Duc Le

Recurrent neural network transducer (RNN-T) is an end-to-end speech recognition framework converting input acoustic frames into a character sequence. The state-of-the-art encoder network for RNN-T is the Conformer, which can effectively…

Audio and Speech Processing · Electrical Eng. & Systems 2022-06-20 Juntae Kim , Jeehye Lee

Sequence transducers, such as the RNN-T and the Conformer-T, are one of the most promising models of end-to-end speech recognition, especially in streaming scenarios where both latency and accuracy are important. Although various methods,…

Audio and Speech Processing · Electrical Eng. & Systems 2022-11-07 Yusuke Shinohara , Shinji Watanabe

We explore options to use Transformer networks in neural transducer for end-to-end speech recognition. Transformer networks use self-attention for sequence modeling and comes with advantages in parallel computation and capturing contexts.…

Audio and Speech Processing · Electrical Eng. & Systems 2019-10-30 Ching-Feng Yeh , Jay Mahadeokar , Kaustubh Kalgaonkar , Yongqiang Wang , Duc Le , Mahaveer Jain , Kjell Schubert , Christian Fuegen , Michael L. Seltzer

Speech recognition on smart devices is challenging owing to the small memory footprint. Hence small size ASR models are desirable. With the use of popular transducer-based models, it has become possible to practically deploy streaming…

Audio and Speech Processing · Electrical Eng. & Systems 2022-01-11 Nauman Dawalatabad , Tushar Vatsal , Ashutosh Gupta , Sungsoo Kim , Shatrughan Singh , Dhananjaya Gowda , Chanwoo Kim

We present Bifocal RNN-T, a new variant of the Recurrent Neural Network Transducer (RNN-T) architecture designed for improved inference time latency on speech recognition tasks. The architecture enables a dynamic pivot for its runtime…

Audio and Speech Processing · Electrical Eng. & Systems 2021-08-05 Jonathan Macoskey , Grant P. Strimel , Ariya Rastrow

Neural transducer-based systems such as RNN Transducers (RNN-T) for automatic speech recognition (ASR) blend the individual components of a traditional hybrid ASR systems (acoustic model, language model, punctuation model, inverse text…

Adaption of end-to-end speech recognition systems to new tasks is known to be challenging. A number of solutions have been proposed which apply external language models with various fusion methods, possibly with a combination of two-pass…

Computation and Language · Computer Science 2021-06-10 Janne Pylkkönen , Antti Ukkonen , Juho Kilpikoski , Samu Tamminen , Hannes Heikinheimo