End-to-End Neural Segmental Models for Speech Recognition | IEEE Journals & Magazine | IEEE Xplore