Building Competitive Direct Acoustics-to-Word Models for English Conversational Speech Recognition | IEEE Conference Publication | IEEE Xplore