Modeling Long Context for Task-Oriented Dialogue State Generation
This work addresses a specific bottleneck in dialogue systems for improving accuracy in multi-turn conversations, though it is incremental as it builds on an existing baseline.
The paper tackles the problem of performance degradation in task-oriented dialogue state generation with long context sequences, achieving a 7.03% relative improvement over the baseline and setting a new state-of-the-art joint goal accuracy of 52.04% on the MultiWOZ 2.0 dataset.
Based on the recently proposed transferable dialogue state generator (TRADE) that predicts dialogue states from utterance-concatenated dialogue context, we propose a multi-task learning model with a simple yet effective utterance tagging technique and a bidirectional language model as an auxiliary task for task-oriented dialogue state generation. By enabling the model to learn a better representation of the long dialogue context, our approaches attempt to solve the problem that the performance of the baseline significantly drops when the input dialogue context sequence is long. In our experiments, our proposed model achieves a 7.03% relative improvement over the baseline, establishing a new state-of-the-art joint goal accuracy of 52.04% on the MultiWOZ 2.0 dataset.