Generative knowledge question-answering method based on representation learning and multi-layer coverage mechanism

Abstract

The invention relates to a generative knowledge question-answering method based on representation learning and a multi-layer coverage mechanism, and belongs to the technical field of artificial intelligence and natural language processing. As for a knowledge base for the knowledge question-answering system, the fact cannot be effectively expressed, so that the correct answer searching capability is reduced. A model in the generative question-answering task sinks into a certain mode and cannot jump out; or the generated vocabularies are repeatedly generated in a certain mode so that the answerreadability is reduced. The method comprises the following steps: firstly, establishing the generative knowledge question-answering model, using a Seq2Seq framework, combining an attention mechanism,a CopyNet model, a GenQA model and a Coverage coverage mechanism, analyzing a question through an encoder, querying information in a knowledge base, and using a decoder to generate an answer. Under agiven scene, a complete sentence can be generated, answers contain correct knowledge, the generated answers have fluency, consistency and correctness, and a good effect is achieved in a classical knowledge question and answer data set, a limited field and a question and answer data set of an open field.

Description

technical field [0001] The invention relates to a generative knowledge question answering method, in particular to a generative knowledge question answering method based on representation learning and a multi-layer covering mechanism, and belongs to the technical field of artificial intelligence and natural language processing. Background technique [0002] Question Answering System (Question Answering System, QA) is an advanced form of information retrieval system, it can use accurate and concise natural language to answer questions raised by users in natural language. The need to obtain information. Question answering system is a research direction that has attracted much attention and has broad development prospects in the field of artificial intelligence and natural language processing. [0003] The task of the knowledge question answering system is to directly search and infer the matching answer in the knowledge base according to the semantics of the user's question. ...

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to overcome the defects of the prior art. In order to solve the knowledge base of the knowledge question answering system, facts cannot be effectively represented, which reduces the ability to find the correct answer, and the model in the generative question answering task often falls into a certain mode and cannot jump out. , or repeating the generated vocabulary in a certain mode, resulting in a technical problem that the readability of the answer is reduced. A generative knowledge question answering method based on representation learning and multi-layer coverage mechanism is proposed.

Images