Grounding Dialogue Systems via Knowledge Graph Aware Decoding with Pre-trained Transformers

Generating knowledge grounded responses in both goal and non-goal oriented\ndialogue systems is an important research challenge. Knowledge Graphs (KG) can\nbe viewed as an abstraction of the real world, which can potentially facilitate\na dialogue system to produce knowledge grounded responses. However, integrating\nKGs into the dialogue generation process in an end-to-end manner is a\nnon-trivial task. This paper proposes a novel architecture for integrating KGs\ninto the response generation process by training a BERT model that learns to\nanswer using the elements of the KG (entities and relations) in a multi-task,\nend-to-end setting. The k-hop subgraph of the KG is incorporated into the model\nduring training and inference using Graph Laplacian. Empirical evaluation\nsuggests that the model achieves better knowledge groundedness (measured via\nEntity F1 score) compared to other state-of-the-art models for both goal and\nnon-goal oriented dialogues.\n