Result Details
Sequence Summarizing Neural Network for Speaker Adaptation
Watanabe Shinji, FIT (FIT)
Žmolíková Kateřina, Ing., Ph.D.
Karafiát Martin, Ing., Ph.D., DCGM (FIT)
Burget Lukáš, doc. Ing., Ph.D., DCGM (FIT)
Černocký Jan, prof. Dr. Ing., DCGM (FIT)
In this paper, we propose a DNN adaptation technique, where the i-vector extractor is replaced by a Sequence Summarizing Neural Network (SSNN). Similarly to i-vector extractor, the SSNN produces a "summary vector", representing an acoustic summary of an utterance. Such vector is then appended to the input of main network, while both networks are trained together optimizing single loss function. Both the i-vector and SSNN speaker adaptation methods are compared on AMI meeting data. The results show comparable performance of both techniques on FBANK system with frameclassification training. Moreover, appending both the i-vector and "summary vector" to the FBANK features leads to additional improvement comparable to the performance of FMLLR adapted DNN system.
DNN, adaptation, i-vector, sequence summary,SSNN
In this paper, we proposed an alternative method to produce DNN adaptation vectors similar to i-vectors. The vectors are computed by the Sequence Summarizing Neural Network and characterize the acoustics in an utterance.
@inproceedings{BUT130964,
author="Karel {Veselý} and Shinji {Watanabe} and Kateřina {Žmolíková} and Martin {Karafiát} and Lukáš {Burget} and Jan {Černocký}",
title="Sequence Summarizing Neural Network for Speaker Adaptation",
booktitle="Proceedings of the 41th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2016), 2016",
year="2016",
pages="5315--5319",
publisher="IEEE Signal Processing Society",
address="Shanghai",
doi="10.1109/ICASSP.2016.7472692",
isbn="978-1-4799-9988-0",
url="https://www.fit.vut.cz/research/publication/11145/"
}Meeting Assistant (MINT), TAČR, Program aplikovaného výzkumu a experimentálního vývoje ALFA, TA04011311, start: 2014-10-01, end: 2017-12-31, completed