Low-Resource Knowledge-Grounded Dialogue Generation
This addresses the challenge of building intelligent conversational agents that can respond with knowledge when training data is scarce, representing an incremental improvement in low-resource settings.
The paper tackles the problem of knowledge-grounded dialogue generation with limited training data by proposing a disentangled response decoder that isolates parameters dependent on such dialogues, allowing most of the model to be trained on abundant ungrounded dialogues and documents. With only 1/8 of the training data, the model achieves state-of-the-art performance and generalizes well on out-of-domain knowledge.
Responding with knowledge has been recognized as an important capability for an intelligent conversational agent. Yet knowledge-grounded dialogues, as training data for learning such a response generation model, are difficult to obtain. Motivated by the challenge in practice, we consider knowledge-grounded dialogue generation under a natural assumption that only limited training examples are available. In such a low-resource setting, we devise a disentangled response decoder in order to isolate parameters that depend on knowledge-grounded dialogues from the entire generation model. By this means, the major part of the model can be learned from a large number of ungrounded dialogues and unstructured documents, while the remaining small parameters can be well fitted using the limited training examples. Evaluation results on two benchmarks indicate that with only 1/8 training data, our model can achieve the state-of-the-art performance and generalize well on out-of-domain knowledge.