Framework Switching of Speaker Overlap Detection System (화자 겹침 검출 시스템의 프레임워크 전환 연구)

Hoinam Kim (김회남); Jisu Park (박지수); Shin Cha (차신); Kyung A Son (손경아); Yun Young-Sun (윤영선); Jeon Gue Park (박전규)

doi:10.29056/jsav.2021.06.13

Framework Switching of Speaker Overlap Detection System

Journal of Software Assessment and Valuation
Abbr : JSAV
2021, 17(1), pp.101-113
DOI : 10.29056/jsav.2021.06.13
Publisher : Korea Software Assessment and Valuation Society
Research Area : Engineering > Computer Science
Received : May 29, 2021
Accepted : June 20, 2021
Published : June 30, 2021

Hoinam Kim ¹, Jisu Park ¹, Shin Cha ¹, Kyung A Son ², Yun Young-Sun ¹, Jeon Gue Park ³

¹한남대학교
²UNIST U교육혁신센터
³한국전자통신연구원

Accredited

ABSTRACT

In this paper, we introduce a speaker overlap system and look at the process of converting the existed system on the specific framework of artificial intelligence. Speaker overlap is when two or more speakers speak at the same time during a conversation, and can lead to performance degradation in the fields of speech recognition or speaker recognition, and a lot of research is being conducted because it can prevent performance degradation. Recently, as application of artificial intelligence is increasing, there is a demand for switching between artificial intelligence frameworks. However, when switching frameworks, performance degradation is observed due to the unique characteristics of each framework, making it difficult to switch frameworks. In this paper, the process of converting the speaker overlap detection system based on the Keras framework to the pytorch-based system is explained and considers components. As a result of the framework switching, the pytorch-based system showed better performance than the existing Keras-based speaker overlap detection system, so it can be said that it is valuable as a fundamental study on systematic framework conversion.

KEYWORDS

speaker overlapping, machine learning framework switch, adam optimizer, learning rate schedule

Citation status

* References for papers published after 2023 are currently being built.

[confproc] Reynolds, Douglas A. / 2005 / Approaches and applications of audio diarization / IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'05) / IEEE 5

[confproc] Bergstra, J. / 2010 / Theano: a CPU and GPU math expression compiler / Proceedings of the Python for scientific computing conference (SciPy) 4(3) : 1~7

[confproc] Abadi, M. / 2016 / Tensorflow: A system for large-scale machine learning / 12th USENIX conference on Operating Systems Design and Implementation : 265~283

[book] Ketkar, N. / 2017 / Deep learning with Python / Apress : 97~111

[confproc] Jia, Y. / 2014 / Caffe: Convolutional architecture for fast feature embedding / Proceedings of the 22nd ACM international conference on Multimedia : 675~678

[other] Paszke, A. / 2019 / Pytorch: An imperative style, high-performance deep learning library / arXiv preprint arXiv:1912.01703

[confproc] Adam, A. G. / 2002 / A new speaker change detection method for two-speaker segmentation / 202 IEEE International Conference on Acoustics, Speech, and Signal Processing 4 : 4~3908

[confproc] Bullock, L. / 2020 / Overlap-aware diarization: Resegmentation using neural end-to-end overlapped speech detection / ICASSP 2020-2020 IEEE Inter-national Conference on Acoustics, Speech and Signal Processing / IEEE : 7114~7118

[web] Garofolo, John S. / 1993 / TIMIT Acoustic-Phonetic Continuous Speech Corpus, LDC93S1 / Linguistic Data Consortium / https://doi.org/10.35111/17gk-bn40

[confproc] G. Gravier / 2012 / The ETAPE corpus for the evaluation of speech-based TV content processing in the French language / Proceedins of the 8th LREC / ELRA : 114~118

[confproc] Sajjan, N. / 2018 / Leveraging LSTM models for overlap detection in multi-party meetings / 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) / IEEE : 5249~5253

[confproc] Andrei, V. / 2017 / Detecting Overlapped Speech on Short Timeframes Using Deep Learning / INTERSPEECH : 1198~1202

[confproc] Kazimirova, E. / 2018 / Automatic Detection of Multi-speaker Fragments with High Time Resolution / INTERSPEECH : 1388~1392

[other] Bahrampour, S. / 2016 / Comparative Study of Deep Learning Software Frameworks / arXiv:1511.06435v3

[web] / 2021 / Effective Tensorflow 2.0 Guide / https://www.tensorflow.org/guide/effective_tf2?hl=en

[other] Ruder, Sebastian / 2016 / An overview of gradient descent optimization algorithms / arXiv preprint arXiv:1609.04747

[confproc] Bottou L. / 2010 / Large-Scale Machine Learning with Stochastic Gradient Descent / Proceedings of COMPSTAT’2010

[confproc] Ilya Sutskever / 2013 / On the importance of initialization and momentum in deep learning / Proceedings of the 30th International Conference on Machine Learning, PMLR 28(3) : 1139~1147

[confproc] Z. Zhang / 2018 / Improved Adam Optimizer for Deep Neural Networks / 2018 IEEE/ACM 26th International Symposium on Quality of Service (IWQoS) : 1~2

[other] Aitor Lewkowycz / 2021 / How to decay your learning rate / arXiv:2103.12682v1

[confproc] D. Snyder / 2018 / X-Vectors:Robust DNN Embeddings for Speaker Recognition / 2018 IEEE International Conference on Acoustics, Speech and Signal Processing : 5329~5333

[web] V. Bushaev / 2021 / Adam — latest trends in deep learning optimization / https://towardsdatascience.com/adam-latest-trends-in-deep-learning-optimization-6be9a291375c

[web] / 2021 / Suboptimal convergence when compared with TensorFlow model / https://discuss.pytorch.org/t/suboptimal-convergence-when-compared-with-tensorflow-model/5099/38

KJCKorea
Journal Central

Journal of Software Assessment and Valuation 2023 KCI Impact Factor : 0.35

Framework Switching of Speaker Overlap Detection System

ABSTRACT

KEYWORDS

Citation status

* References for papers published after 2023 are currently being built.

Journal of Software Assessment and Valuation 2023 KCI Impact Factor : 0.35

Framework Switching of Speaker Overlap Detection System

ABSTRACT

KEYWORDS

Statistics

Tools

Issue List

Citation status

KCI Citation Counts (0)

REFERENCES (23) * References for papers published after 2023 are currently being built.

Search PDF

Citation

* References for papers published after 2023 are currently being built.