Timezone: »

A fully differentiable beam search decoder
Ronan Collobert · Awni Hannun · Gabriel Synnaeve

Thu Jun 13 06:30 PM -- 09:00 PM (PDT) @ Pacific Ballroom #226

We introduce a new beam search decoder that is fully differentiable, making it possible to optimize at training time through the inference procedure. Our decoder allows us to combine models which operate at different granularities (e.g. acoustic and language models). It can be used when target sequences are not aligned to input sequences by considering all possible alignments between the two. We demonstrate our approach scales by applying it to speech recognition, jointly training acoustic and word-level language models. The system is end-to-end, with gradients flowing through the whole architecture from the word-level transcriptions. Recent research efforts have shown that deep neural networks with attention-based mechanisms can successfully train an acoustic model from the final transcription, while implicitly learning a language model. Instead, we show that it is possible to discriminatively train an acoustic model jointly with an \emph{explicit} and possibly pre-trained language model.

Author Information

Ronan Collobert (Facebook AI Research)
Awni Hannun (Facebook AI Research)
Gabriel Synnaeve (Facebook AI Research)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors