Dialogue generation method fusing knowledge

A knowledge and knowledge map technology, applied in the computer field, can solve problems such as dialogue system problems that cannot be solved well, and achieve the effect of solving topic concept drift and expansion

Active Publication Date: 2021-01-15
中科厦门数据智能研究院
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this simple and rude approach cannot solve the problems faced by the dialogue system very well.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Dialogue generation method fusing knowledge
  • Dialogue generation method fusing knowledge
  • Dialogue generation method fusing knowledge

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0031] A method for generating dialogues with fusion of knowledge, comprising the following steps:

[0032] S1. Build a knowledge map;

[0033] The knowledge graph of this embodiment adopts the open source knowledge graph Freebase, and one or more of open source knowledge graphs such as OpenCyc, WordNet, Freebase, Dbpedia, and ConceptNet can be selected, or domain knowledge graphs constructed based on business scenarios, and the method of construction Use conventional means of building knowledge graphs, such as named entity recognition, entity disambiguation, and entity unification.

[0034] S2, build dialogue model, described dialogue model is made up of encoder and decoder, and described encoder comprises word encoding, knowledge encoding and double-hop entity encoding;

[0035] Such as figure 1 As shown, the specific steps for building a dialogue model are as follows:

[0036] S21, word coding: the word coding is to utilize the Bert model of google open source to extract...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a dialogue generation method fusing knowledge. The dialogue generation method comprises the following steps that S1, a knowledge graph is constructed; and S2, a dialogue modelis constructed, the dialogue model is composed of an encoder and a decoder, and the construction process comprises word encoding, knowledge encoding, double-hop entity encoding, weighted merging and decoding. According to the method, graph coding and a graph attention mechanism are introduced to carry out double-hop entity coding, and entity semantics in dialogue are better captured based on the relationship between adjacent entities; and meanwhile, in combination with concept knowledge involved in dialogue surrounded by the knowledge graph, a more reasonable reply rich in information is given, and the problem of topic concept drifting and extension in the current dialogue process is solved.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a dialog generation method for merging knowledge. Background technique [0002] With the development of language model and natural language generation technology, data-driven end-to-end dialogue generation technology has become possible. However, end-to-end generated dialogues often produce some meaningless, useless and off-topic replies, such as some safe answers: "I don't know yet", as well as the drift and expansion of topic concepts, which bring discomfort to the dialogue. nice experience. Since language comprehension is closely related to expression and knowledge, for such problems, it is common practice to add external knowledge information to the input of the model. The existing practice is to directly embed the external knowledge text into the vector space, obtain the word vector representation of the external knowledge and input it into the dialogue model. However, ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/332G06F16/36G06F40/126G06F40/205G06F40/284G06F40/295
CPCG06F16/367G06F16/3329G06F40/126G06F40/284G06F40/295G06F40/205Y02D10/00
Inventor 倪斌庄福振安竹林程坦徐勇军陆晓蕾马涛童逸琦
Owner 中科厦门数据智能研究院
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products