Publication Details
Out-of-Vocabulary Word Recovery Using FST-Based Subword Unit Clustering in a Hybrid ASR System
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Out-of-vocabulary Words, Robust ASR
The paper presents a new approach to extracting useful informationfrom out-of-vocabulary (OOV) speech regions inASR system output. The system makes use of a hybrid decodingnetwork with both words and sub-word units. In thedecoded lattices, candidates for OOV regions are identifiedas sub-graphs of sub-word units. To facilitate OOV word recovery,we search for recurring OOVs by clustering the detectedcandidate OOVs. The metrics for clustering is basedon a comparison of the sub-graphs corresponding to the OOVcandidates. The proposed method discovers repeating outof-vocabulary words and finds their graphemic representationmore robustly than more conventional techniques taking intoaccount only one best sub-word string hypotheses.
@inproceedings{BUT155047,
author="Ekaterina {Egorova} and Lukáš {Burget}",
title="Out-of-Vocabulary Word Recovery Using FST-Based Subword Unit Clustering in a Hybrid ASR System",
booktitle="Proceedings of ICASSP 2018",
year="2018",
pages="5919--5923",
publisher="IEEE Signal Processing Society",
address="Calgary",
doi="10.1109/ICASSP.2018.8462221",
isbn="978-1-5386-4658-8",
url="https://www.fit.vut.cz/research/publication/11725/"
}