Unsupervised Learning of Interpretable Dialog Models
This addresses the challenge of making dialog models interpretable for developers and users, representing an incremental improvement by combining unsupervised learning with discrete state tracking.
The paper tackles the problem of interpretability in deep learning-based dialog models by proposing the Latent State Tracking Network (LSTN), which learns discrete latent variables in an unsupervised manner, achieving interpretability without significant performance loss compared to end-to-end approaches.
Recently several deep learning based models have been proposed for end-to-end learning of dialogs. While these models can be trained from data without the need for any additional annotations, it is hard to interpret them. On the other hand, there exist traditional state based dialog systems, where the states of the dialog are discrete and hence easy to interpret. However these states need to be handcrafted and annotated in the data. To achieve the best of both worlds, we propose Latent State Tracking Network (LSTN) using which we learn an interpretable model in unsupervised manner. The model defines a discrete latent variable at each turn of the conversation which can take a finite set of values. Since these discrete variables are not present in the training data, we use EM algorithm to train our model in unsupervised manner. In the experiments, we show that LSTN can help achieve interpretability in dialog models without much decrease in performance compared to end-to-end approaches.