EP3707703A1 - TRAINING AND/OR USING A LANGUAGE SELECTION MODEL FOR AUTOMATICALLY DETERMINING LANGUAGE FOR SPEECH RECOGNITION OF SPOKEN UTTERANCE