Question-Answering Reasoning Algorithm based on Large Language Model

In the face of the growing demand, the question answering system needs strong reasoning ability in order to better provide users with satisfactory answers. Large language model is a large-scale pre-training model based on deep learning. By learning massive text data, it can master the grammar, semantics and context information of natural language, understand the intention of users to ask questions, and retrieve answers related to questions from a large number of texts, which provides strong support for the construction of intelligent question answering system. This paper studies the question answering inference algorithm based on large language model, and explores the user-centered mechanism to provide a solution for constructing a more scientific and applicable question answering inference system. Based on the Transformer model and BERT model, this paper focuses on the integrated Transformer encoder structure in BERT, and captures the complex interdependencies between words through SAM.