PHMM based asynchronous acoustic model for Chinese large vocabulary continuous speech recognition | IEEE Conference Publication | IEEE Xplore