Disentangling Language and Knowledge in Task-Oriented Dialogs
This addresses the issue of maintaining performance in real-world dialog applications like booking systems when knowledge updates occur, representing a strong specific gain rather than a foundational breakthrough.
The paper tackles the problem of task-oriented dialog systems breaking down when knowledge bases change, by proposing BoSsNet with a Bag-of-Sequences memory to disentangle language and knowledge learning, resulting in over 10% improvements on bAbI OOV test sets and robustness to KB modifications.
The Knowledge Base (KB) used for real-world applications, such as booking a movie or restaurant reservation, keeps changing over time. End-to-end neural networks trained for these task-oriented dialogs are expected to be immune to any changes in the KB. However, existing approaches breakdown when asked to handle such changes. We propose an encoder-decoder architecture (BoSsNet) with a novel Bag-of-Sequences (BoSs) memory, which facilitates the disentangled learning of the response's language model and its knowledge incorporation. Consequently, the KB can be modified with new knowledge without a drop in interpretability. We find that BoSsNet outperforms state-of-the-art models, with considerable improvements (> 10\%) on bAbI OOV test sets and other human-human datasets. We also systematically modify existing datasets to measure disentanglement and show BoSsNet to be robust to KB modifications.