Speech silicon AM: an FPGA-based acoustic modeling pipeline for hidden Markov model based speech recognition | IEEE Conference Publication | IEEE Xplore