Recent Advances in Retrieval-Augmented Text Generation
90-Minute Tutorial @ IJCAI Conference 2022.
Overview
Recently, retrieval-augmented text generation has achieved state-of-the-art performance in many NLP tasks and attracted increasing attention of the computational linguistics community. Since retrieval-augmented generation is distributed in many sub-domains of information retrieval and text generation, it increases the difficulty for newcomers to get started. Therefore, this tutorial aims to introduce recent advances in retrieval-augmented text generation. It firstly highlights the generic paradigm of retrieval-augmented generation, and then it reviews notable approaches according to different tasks including dialogue generation, machine translation, and other generation tasks. Finally, it points out some limitations and shortcomings for recent approaches to facilitate future research.
