Selection and generation of faithful responses. This straightforward and intuitive inference procedure of SPIĭirectly queries the response generation model, allowing for accurate knowledge Require the inference network or assume a simple geometry of the posteriorĭistribution. Sampling from the posterior distribution. In this paper, we present anĮnd-to-end learning framework, termed Sequential Posterior Inference (SPI),Ĭapable of selecting knowledge and generating dialogues by approximately Generation by employing an inference network. Variational method to jointly optimize knowledge selection and response Inherent correlation between these two tasks, or leverage conditional Knowledge selection and response generation separately, and may overlook the Common strategies either adopt a two-step paradigm, which optimizes
Download a PDF of the paper titled Diverse and Faithful Knowledge-Grounded Dialogue Generation via Sequential Posterior Inference, by Yan Xu and 6 other authors Download PDF Abstract: The capability to generate responses with diversity and faithfulness usingįactual knowledge is paramount for creating a human-like, trustworthy dialogue
0 kommentar(er)
