TY - GEN
T1 - Designing an Effective Metric Learning Pipeline for Speaker Diarization
AU - Narayanaswamy, Vivek Sivaraman
AU - Thiagarajan, Jayaraman J.
AU - Song, Huan
AU - Spanias, Andreas
N1 - Funding Information:
This work was supported in part by the ASU SenSIP center, Arizona State University. Portions of this work were performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-PROC-767885.
Publisher Copyright:
© 2019 IEEE.
PY - 2019/5
Y1 - 2019/5
N2 - State-of-the-art speaker diarization systems utilize knowledge from external data, in the form of a pre-trained distance metric, to effectively determine relative speaker identities to unseen data. However, much of recent focus has been on choosing the appropriate feature extractor, ranging from pre-trained i-vectors to representations learned via different sequence modeling architectures (e.g. 1D-CNNs, LSTMs, attention models), while adopting off-the-shelf metric learning solutions. In this paper, we argue that, regardless of the feature extractor, it is crucial to carefully design a metric learning pipeline, namely the loss function, the sampling strategy and the discriminative margin parameter, for building robust diarization systems. Furthermore, we propose to adopt a fine-grained validation process to obtain a comprehensive evaluation of the generalization power of metric learning pipelines. To this end, we measure diarization performance across different language speakers, and variations in the number of speakers in a recording. Using empirical studies, we provide interesting insights into the effectiveness of different design choices and make recommendations.
AB - State-of-the-art speaker diarization systems utilize knowledge from external data, in the form of a pre-trained distance metric, to effectively determine relative speaker identities to unseen data. However, much of recent focus has been on choosing the appropriate feature extractor, ranging from pre-trained i-vectors to representations learned via different sequence modeling architectures (e.g. 1D-CNNs, LSTMs, attention models), while adopting off-the-shelf metric learning solutions. In this paper, we argue that, regardless of the feature extractor, it is crucial to carefully design a metric learning pipeline, namely the loss function, the sampling strategy and the discriminative margin parameter, for building robust diarization systems. Furthermore, we propose to adopt a fine-grained validation process to obtain a comprehensive evaluation of the generalization power of metric learning pipelines. To this end, we measure diarization performance across different language speakers, and variations in the number of speakers in a recording. Using empirical studies, we provide interesting insights into the effectiveness of different design choices and make recommendations.
KW - Speaker diarization
KW - attention models
KW - inverse distance weighted sampling
KW - metric learning
UR - http://www.scopus.com/inward/record.url?scp=85068996908&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85068996908&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2019.8682255
DO - 10.1109/ICASSP.2019.8682255
M3 - Conference contribution
AN - SCOPUS:85068996908
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 5806
EP - 5810
BT - 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019
Y2 - 12 May 2019 through 17 May 2019
ER -