Speech Recognition on Code-Switching Among the Chinese Dialects

Proceedings of the Third Workshop on Computational Approaches to Linguistic Code-Switching

“…The Language Identification (LID) based approach is to identify the language boundaries and subsequently use an appropriate monolingual ASR system to recognize monolingual fragments (Chan et al, 2004) or run multiple recognizers in parallel with an LID system (Ahmed and Tan, 2012;). Another approach is to train an AM on bilingual data (Lyu et al, 2006; or to use one of the monolingual AMs (Bhuvanagirir and Kopparapu, 2012) or to pool the existing monolingual AMs by sharing phones belonging to both languages. Yeh et al (Yeh and Lee, 2015) tackle the problem of code-switching in which a speaker speaks mainly in one language, leading to an imbalance in the amount of data available in the two languages, with cross-lingual data sharing approaches.…”

“…(2)Employing a shared phone set to build acoustic models for mixed speech with standard language models trained on codeswitched text (Imseng et al, 2011;Li et al, 2011;Bhuvanagiri and Kopparapu, 2010;Yeh et al, 2010). (3) Training Acoustic or Language models on monolingual data in both languages with little or no code-switched data (Lyu et al, 2006;Bhuvanagirir and Kopparapu, 2012;Yeh and Lee, 2015). We attempt to approach this (1) Approaches that use DNNs as feature extractors followed by separate classifiers to predict the identity of the language (Jiang et al, 2014;Matejka et al, 2014;Song et al, 2013) and 2Approaches that employ DNNs to directly predict the language ID (Richardson et al, 2015b,a;Lopez-Moreno et al, 2014).…”

Mixed Language Speech Recognition without Explicit Identification of Language

“…Clearly a poor performance by any one of the three blocks affects the overall performance of the mult i pass based ML-A SR system. The one pass framework [3] avoids the drawback of mult i pass system by building a PL, AM and LM to encompass both the languages in the mixed language. The acoustic model for mixed language is an AM generated for the co mbined phoneme set of the languages in the mixed language.…”

“…There are two approaches reported in literature. One being mult i pass fra mework [4] and other is the one pass framework [3]. Ho wever, mult ilingual speech recognition is another area of research which has close relationship with ML-ASR.…”

“…Clearly, th is kind of system cannot work when people use mixed language speech, even if one knew the mix o f languages in use, because the language shift is very frequent. Recently there has been increased interest in mixed language recognition (for example [3], [4], [19]) however the work has been restricted to a mix of Mandarin and Taiwanese. Mixed language speech recognition is in its nascent stages of research and to the best of our knowledge there is no work reported in literature for India specific language mix.…”