Direct speech-to-speech translation with a sequence-to-sequence model
This addresses the challenge of speech-to-speech translation for users needing real-time communication, but it is incremental as it underperforms existing methods.
The authors tackled the problem of translating speech directly between languages without intermediate text, using an end-to-end sequence-to-sequence model that maps spectrograms and can mimic the source speaker's voice. They found the model slightly underperformed a baseline cascade on Spanish-to-English datasets, showing feasibility but not superiority.
We present an attention-based sequence-to-sequence neural network which can directly translate speech from one language into speech in another language, without relying on an intermediate text representation. The network is trained end-to-end, learning to map speech spectrograms into target spectrograms in another language, corresponding to the translated content (in a different canonical voice). We further demonstrate the ability to synthesize translated speech using the voice of the source speaker. We conduct experiments on two Spanish-to-English speech translation datasets, and find that the proposed model slightly underperforms a baseline cascade of a direct speech-to-text translation model and a text-to-speech synthesis model, demonstrating the feasibility of the approach on this very challenging task.