A comparison of linguistically and statistically enhanced models for speech-to-speech machine translation

20500

Knowledge Triples