Towards robustness to fast speech in ASR | IEEE Conference Publication | IEEE Xplore