Multitask Learning Model with Text and Speech Representation for Fine-Grained Speech Scoring | IEEE Conference Publication | IEEE Xplore