Abstract
Dialogue state tracking, which estimates user goals and requests given thedialogue context, is an essential part of task-oriented dialogue systems. Inthis paper, we propose the Global-Locally Self-Attentive Dialogue State Tracker(GLAD), which learns representations of the user utterance and previous systemactions with global-local modules. Our model uses global modules to shareparameters between estimators for different types (called slots) of dialoguestates, and uses local modules to learn slot-specific features. We show thatthis significantly improves tracking of rare states and achievesstate-of-the-art performance on the WoZ and DSTC2 state tracking tasks. GLADobtains 88.1% joint goal accuracy and 97.1% request accuracy on WoZ,outperforming prior work by 3.7% and 5.5%. On DSTC2, our model obtains 74.5%joint goal accuracy and 97.5% request accuracy, outperforming prior work by1.1% and 1.0%.