Acoustic modeling with a shared phoneme set for multilingual speech recognition without code-switching | IEEE Conference Publication | IEEE Xplore