US20200082809A1 - USING RECURRENT NEURAL NETWORK FOR PARTITIONING OF AUDIO DATA INTO SEGMENTS THAT EACH CORRESPOND TO A SPEECH FEATURE CLUSTER IDENTIFIER