Revisiting joint decoding based multi-talker speech recognition with DNN acoustic model

Revisiting joint decoding based multi-talker speech recognition with DNN acoustic model

Stať ve sborníku v databázi WoS či Scopus

In typical multi-talker speech recognition systems, a neural network-based acoustic model predicts senone state posteriors for each speaker. These are later used by a single-talker decoder which is applied on each speaker-specific output stream separately. In this work, we argue that such a scheme is sub-optimal and propose a principled solution that decodes all speakers jointly. We modify the acoustic model to predict joint state posteriors for all speakers, enabling the network to express uncertainty about the attribution of parts of the speech signal to the speakers. We employ a joint decoder that can make use of this uncertainty together with higher-level language information. For this, we revisit decoding algorithms used in factorial generative models in early multi-talker speech recognition systems. In contrast with these early works, we replace the GMM acoustic model with DNN, which provides greater modeling power and simplifies part of the inference. We demonstrate the advantage of joint decoding in proof of concept experiments on a mixed-TIDIGITS dataset.

In typical multi-talker speech recognition systems, a neural network-based acoustic model predicts senone state posteriors for each speaker. These are later used by a single-talker decoder which is applied on each speaker-specific output stream separately. In this work, we argue that such a scheme is sub-optimal and propose a principled solution that decodes all speakers jointly. We modify the acoustic model to predict joint state posteriors for all speakers, enabling the network to express uncertainty about the attribution of parts of the speech signal to the speakers. We employ a joint decoder that can make use of this uncertainty together with higher-level language information. For this, we revisit decoding algorithms used in factorial generative models in early multi-talker speech recognition systems. In contrast with these early works, we replace the GMM acoustic model with DNN, which provides greater modeling power and simplifies part of the inference. We demonstrate the advantage of joint decoding in proof of concept experiments on a mixed-TIDIGITS dataset.

Multi-talker speech recognition, Permutation invariant training, Factorial Hidden Markov models

Multi-talker speech recognition, Permutation invariant training, Factorial Hidden Markov models

KOCOUR, M.; ŽMOLÍKOVÁ, K.; ONDEL YANG, L.; ŠVEC, J.; DELCROIX, M.; OCHIAI, T.; BURGET, L.; ČERNOCKÝ, J.

2023

18.09.2022

International Speech Communication Association

Incheon

Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

1990-9772

Proceedings of Interspeech

9

Francouzská republika

4955

4959

5

https://www.isca-speech.org/archive/pdfs/interspeech_2022/kocour22_interspeech.pdf

kocour22_interspeech2022_revisiting