Autoregressive Entity Generation for End-to-End Task-Oriented Dialog
This work addresses entity consistency issues in end-to-end task-oriented dialog systems, which is an incremental improvement for applications like customer service or virtual assistants.
The paper tackled the problem of generating conflicting entity information in task-oriented dialog systems by proposing an autoregressive entity generation method with a trie constraint and logit concatenation, resulting in more high-quality and entity-consistent responses as shown in experiments on MultiWOZ 2.1 and CAMREST datasets.
Task-oriented dialog (TOD) systems often require interaction with an external knowledge base to retrieve necessary entity (e.g., restaurant) information to support the response generation. Most current end-to-end TOD systems either retrieve the KB information explicitly or embed it into model parameters for implicit access.~While the former approach demands scanning the KB at each turn of response generation, which is inefficient when the KB scales up, the latter approach shows higher flexibility and efficiency. In either approach, the systems may generate a response with conflicting entity information. To address this issue, we propose to generate the entity autoregressively first and leverage it to guide the response generation in an end-to-end system. To ensure entity consistency, we impose a trie constraint on entity generation. We also introduce a logit concatenation strategy to facilitate gradient backpropagation for end-to-end training. Experiments on MultiWOZ 2.1 single and CAMREST show that our system can generate more high-quality and entity-consistent responses.