Use of higher level linguistic structure in acoustic modeling for speech recognition | IEEE Conference Publication | IEEE Xplore