Learning Acoustic Frame Labeling for Phoneme Segmentation with Regularized Attention Mechanism | IEEE Conference Publication | IEEE Xplore