Retrieval Augmentation Reduces Hallucination in Conversation
This addresses the issue of knowledge hallucination in chatbots, which is critical for improving reliability in conversational AI, though it builds on existing retrieval methods from open-domain QA.
The paper tackled the problem of factual incorrectness and hallucination in dialogue models by applying neural-retrieval-in-the-loop architectures to knowledge-grounded dialogue, resulting in state-of-the-art performance on two tasks and a substantial reduction in hallucination as verified by human evaluations.
Despite showing increasingly human-like conversational abilities, state-of-the-art dialogue models often suffer from factual incorrectness and hallucination of knowledge (Roller et al., 2020). In this work we explore the use of neural-retrieval-in-the-loop architectures - recently shown to be effective in open-domain QA (Lewis et al., 2020b; Izacard and Grave, 2020) - for knowledge-grounded dialogue, a task that is arguably more challenging as it requires querying based on complex multi-turn dialogue context and generating conversationally coherent responses. We study various types of architectures with multiple components - retrievers, rankers, and encoder-decoders - with the goal of maximizing knowledgeability while retaining conversational ability. We demonstrate that our best models obtain state-of-the-art performance on two knowledge-grounded conversational tasks. The models exhibit open-domain conversational capabilities, generalize effectively to scenarios not within the training data, and, as verified by human evaluations, substantially reduce the well-known problem of knowledge hallucination in state-of-the-art chatbots.