Self-supervised attention flow for dialogue state tracking

摘要

The performance of existing approaches for dialogue state tracking (DST) is often limited by the defi-ciency of labeled datasets, and inefficient utilization of data is also a practical yet tough problem of the DST task. In this paper, we aim to tackle these challenges in a self-supervised manner by introducing an auxiliary pre-training task that learns to pick up the correct dialogue response from a group of candi-dates. Moreover, we propose an attention flow mechanism that is augmented with a soft-threshold func-tion in a dynamic way to better understand the user intent and filter out the redundant information. Extensive experiments on the multi-domain dialogue state tracking dataset MultiWOZ 2.1 demonstrate the effectiveness of our proposed method, and we also show that it is able to adapt to zero/few-shot cases under the proposed self-supervised framework.(c) 2021 Elsevier B.V. All rights reserved.