A method for automatic answer text generation using multi-view information

Through the multi-view automatic answer generation model, the multi-view information is integrated to generate natural, smooth and accurate answers, which solves the problem of insufficient fusion of multi-view information in the existing technology, and improves the fluency and information richness of the answers.

CN114692649BActive Publication Date: 2025-03-21SOUTH CHINA UNIV OF TECH
2 Cites 0 Cited by

Patent Information

Application Number
CN202210286420.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-23
Publication Date
2025-03-21
Estimated Expiration
2042-03-23

AI Technical Summary

Technical Problem

When the existing automatic question and answer system generates answer text, the multi-view information fusion is insufficient, resulting in insufficient fluency and richness of the generated answers.

Method used

The multi-view automatic answer generation model is adopted, and by training the reading comprehension and answer generator, multi-view information is integrated to generate natural, smooth and accurate answers. The model includes an encoder, self-attention layer, interactive attention layer, segmented attention layer, decoder and multi-view fusion device, which can effectively understand and fuse multi-view information.

Benefits of technology

It improves the fluency and information richness of the answers, making the generated answers more smooth, free, flexible and authentic, and is suitable for scenarios such as Q&A community, medical Q&A and customer service automatic answers.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present invention discloses a method for generating automatic answer text using multi-perspective information, including: obtaining basic data; preprocessing the data and building a multi-perspective automatic answer generation model; using the preprocessed data to train the multi-perspective automatic answer generation model to obtain a model with the best generation results; after the user asks a question, collecting relevant information, using the trained multi-perspective automatic answer generation model to generate results, and feeding the results back to the user. The present invention trains a model that can understand multi-perspective information and fuses this information to generate natural, fluent and accurate answers. When a user asks a question, the model is used to automatically generate text to answer the user's question and improve the user's understanding of the target. The method of the present invention has a wide range of uses in scenarios such as question-and-answer communities, medical question-and-answer, and customer service automatic answers.
Need to check novelty before this filing date? Find Prior Art