Re-Translation Strategies For Long Form, Simultaneous, Spoken Language Translation
This addresses the problem of generating stable, real-time translated captions for live audio feeds like lectures, though it is incremental in its use of existing tools.
The paper tackled simultaneous machine translation of long-form speech by adopting a re-translation approach, which achieved very low latency and high final quality but with incremental instability, and improved stability across seven target languages using simple heuristics.
We investigate the problem of simultaneous machine translation of long-form speech content. We target a continuous speech-to-text scenario, generating translated captions for a live audio feed, such as a lecture or play-by-play commentary. As this scenario allows for revisions to our incremental translations, we adopt a re-translation approach to simultaneous translation, where the source is repeatedly translated from scratch as it grows. This approach naturally exhibits very low latency and high final quality, but at the cost of incremental instability as the output is continuously refined. We experiment with a pipeline of industry-grade speech recognition and translation tools, augmented with simple inference heuristics to improve stability. We use TED Talks as a source of multilingual test data, developing our techniques on English-to-German spoken language translation. Our minimalist approach to simultaneous translation allows us to easily scale our final evaluation to six more target languages, dramatically improving incremental stability for all of them.