Robust Large Vocabulary Continuous Speech Recognition using Polynomial Segment Model with Unsupervised Adaptation | IEEE Conference Publication | IEEE Xplore