Zero-Shot Transfer Learning with Synthesized Data for Multi-Domain Dialogue State Tracking
This work addresses the high cost of data acquisition for new domains in dialogue systems, though it is incremental as it builds on existing models like TRADE and SUMBT.
The paper tackles the problem of zero-shot transfer learning for multi-domain dialogue state tracking by using synthesized data from an abstract dialogue model and domain ontology, achieving a 21% average improvement in zero-shot learning state of the art across domains.
Zero-shot transfer learning for multi-domain dialogue state tracking can allow us to handle new domains without incurring the high cost of data acquisition. This paper proposes new zero-short transfer learning technique for dialogue state tracking where the in-domain training data are all synthesized from an abstract dialogue model and the ontology of the domain. We show that data augmentation through synthesized data can improve the accuracy of zero-shot learning for both the TRADE model and the BERT-based SUMBT model on the MultiWOZ 2.1 dataset. We show training with only synthesized in-domain data on the SUMBT model can reach about 2/3 of the accuracy obtained with the full training dataset. We improve the zero-shot learning state of the art on average across domains by 21%.