ASR domain adaptation methods for low-resourced languages: Application to Romanian language

This study investigates the possibility of using statistical machine translation to create domain-specific language resources. We propose a methodology that aims to create a domain-specific automatic speech recognition system for a low-resourced language when in-domain text corpora are available only in a high-resourced language. We evaluate a new semi-supervised method and compare it with previously developed semi-supervised and unsupervised approaches. Moreover, in the effort of creating an out-of-domain language model for Romanian, we introduce and experiment an effective diacritics restoration algorithm.