On semi-supervised LF-MMI training of acoustic models with limited data

Imran Sheikh (1) , Emmanuel Vincent (1) , Irina Illina (1)

1 MULTISPEECH - Speech Modeling for Facilitating Oral-Based Communication

This work investigates semi-supervised training of acoustic models (AM) with the lattice-free maximum mutual information (LF-MMI) objective in practically relevant scenarios with a limited amount of labeled in-domain data. An error detection driven semi-supervised AM training approach is proposed, in which an error detector controls the hypothesized transcriptions or lattices used as LF-MMI training targets on additional unlabeled data. Under this approach, our first method uses a single error-tagged hypothesis whereas our second method uses a modified supervision lattice. These methods are evaluated and compared with existing semi-supervised AM training methods in three different matched or mismatched, limited data setups. Word error recovery rates of 28 to 89% are reported.

Mots clés: lattice-free MMI; semi-supervised training; speech recognition; error detection

Domaines: Informatique et langage [cs.CL]; Apprentissage [cs.LG]

Imran Sheikh, Emmanuel Vincent, Irina Illina. On semi-supervised LF-MMI training of acoustic models with limited data. INTERSPEECH 2020, Oct 2020, Shanghai, China. ⟨hal-02907924⟩