Answer prediction method and device based on graph reasoning model

Abstract

The invention discloses an answer prediction method and device based on a graph reasoning model. The method comprises the following steps of receiving questions and a support document set; screening out irrelevant documents through text slimming, and performing semantic coding on all texts; semantic interaction of various texts and initialization of graph nodes are carried out by using various attention mechanisms; performing multi-hop reasoning by using a GNN information transfer algorithm based on the constructed graph; calculating a score for predicting an answer for each candidate based on the updated graph node representation; and predicting an unknown right entity, namely an answer, according to the score distribution result of the answer. A new graph is provided, multiple types of elements are regarded as graph nodes, and reasoning is more comprehensive. Meanwhile, due to the fact that sentence nodes are adopted, reasoning becomes more accurate and specific, multiple attention mechanisms are fused to carry out multiple semantic representation, the influence of relative correctness between candidates on reasoning is innovatively considered, and answer prediction is more accurate.

Description

technical field [0001] The invention belongs to the technical field of natural language processing in artificial intelligence, and in particular relates to an answer prediction method and device based on a graph reasoning model. Background technique [0002] Machine reading comprehension (MRC) is mainly used to measure the degree of machine understanding of natural language content, which is an important step in the realization of artificial intelligence. Usually, the MRC task will give a document and a question, and the machine needs to select, extract and fuse the key semantic information in it, and try to answer the question related to the document, which is a complex natural language processing task. In recent years, with the rise of the army of MRC researchers, many high-quality data sets have been proposed to evaluate the development level of MRC, such as SQuAD, RACE, etc. Many neural models have been proposed for these tasks (BiDAF, Match-LSTM), and, after some excel...

AI Technical Summary

Problems solved by technology

However, only using entities, sentences or candidates as GNN nodes to achieve reasoning and capture key information is not enough to support accurate reasoning

Of course, there are also some models that use two types of nodes to build graphs, but this is not enough

At the same time, the graph reasoning based on the entity level is too fine, while the reasoning based on the document level is too rough, and the reasoning effect is suppressed.

Images