End-to-end trainable networks try to overcome error propagation, lack of generalization and overall brittleness of traditional modularized task-oriented dialogue system architectures. Most proposed models expand on the sequence-to-sequence architecture. Some of them don’t track belief state, which makes it difficult to interact with ever-changing knowledge bases, while the ones that explicitly track the belief state do it with classifiers. The use of classifiers suffers from the out-of-vocabulary words problem, making these models hard to use in real-world applications with ever-changing knowledge bases. We propose Structured Belief Copy Network (SBCN), a novel end-to-end trainable architecture that allows for interaction with external symbolic knowledge bases and solves the out-of-vocabulary problem at the same time. It explicitly uses the structure of the belief state in its architecture by using different sequential decoders with a copy mechanism for the different informable slots and a multi-label decoder for the requestable slots, and this provides better inductive bias. Through empirical results we show that SBCN achieves state-of-the-art results on dialogue datasets while providing a practical architecture that can be used in real-world applications.
ConvAI @ NeurIPS 2018