End-to-End Joint Learning of Natural Language Understanding and Dialogue Manager
This addresses the issue of noisy NLU outputs affecting dialogue policy in conversational systems, representing an incremental improvement over existing methods.
The paper tackles the problem of error propagation in conversational systems by proposing an end-to-end joint learning model for natural language understanding and dialogue management, which significantly outperforms state-of-the-art pipeline models on DSTC4 dialogues.
Natural language understanding and dialogue policy learning are both essential in conversational systems that predict the next system actions in response to a current user utterance. Conventional approaches aggregate separate models of natural language understanding (NLU) and system action prediction (SAP) as a pipeline that is sensitive to noisy outputs of error-prone NLU. To address the issues, we propose an end-to-end deep recurrent neural network with limited contextual dialogue memory by jointly training NLU and SAP on DSTC4 multi-domain human-human dialogues. Experiments show that our proposed model significantly outperforms the state-of-the-art pipeline models for both NLU and SAP, which indicates that our joint model is capable of mitigating the affects of noisy NLU outputs, and NLU model can be refined by error flows backpropagating from the extra supervised signals of system actions.