Supercharge Your Innovation With Domain-Expert AI Agents!

Personalized dialogue generation method and system based on user dialogue history

A user and history technology, applied in character and pattern recognition, special data processing applications, unstructured text data retrieval, etc., can solve the problems that new users cannot join the model, ignore, and role information cannot be strengthened

Active Publication Date: 2021-01-05
RENMIN UNIVERSITY OF CHINA
View PDF9 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

To a certain extent, this makes the implicit person setting change from a static vector to a dynamic vector, which can be adjusted at different stages of generation. However, the reply generated by means of personalized keywords can only have the significant attribute information of the character. Subtle personal characteristics such as language habits during character dialogue cannot be captured
In addition, at present, the personality vector is mainly acquired through the user ID during training and updated through backpropagation. With the increase of users, the proportion of each user’s historical number in the entire data set will decrease, and the user’s personality vector will increase with time. As the model is updated slowly, new users cannot join the already trained model, and the model needs to be retrained. In practical applications, it is difficult to respond to the user's new history and new users in a timely manner.
At the same time, this method ignores the timing information in the user's dialogue history, and the user's latest role information cannot be strengthened in the personality vector

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
  • Personalized dialogue generation method and system based on user dialogue history
  • Personalized dialogue generation method and system based on user dialogue history
  • Personalized dialogue generation method and system based on user dialogue history

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] This embodiment discloses a method for generating personalized dialogue based on long-short-term memory information, such as figure 1 shown, including the following steps:

[0041] S1 represents the input text and the text of user dialogue history as sentence vectors.

[0042] S2 obtains the user personality vector by encoding the sentence vector, and the user personality vector contains the time sequence information of the sentence vector.

[0043] In this step, the sentence vector is mainly processed by the Seq2Seq model and the attention mechanism. The Seq2Seq model encodes sentence vectors and combines them via an attention mechanism to generate responses for the decoding process.

[0044] A Seq2Seq model usually consists of an encoder and a decoder. The role of the encoder is to represent the input text X, and convert the input text X into a dense vector H=(h 1 , h 2 ,...,h n ). The role of the decoder is to convert this intermediate state vector h n Decode...

Embodiment 2

[0075] Based on the same inventive concept, this embodiment discloses a personalized dialog generation system based on user dialog history, including:

[0076] The sentence vector generation module is used to represent the text of the user dialogue history as a sentence vector;

[0077] A personality vector generation module, used to obtain the user personality vector by encoding the sentence vector, the user personality vector includes the timing information of the sentence vector;

[0078] Model generation mode, which is used to generate a personalized dialogue model according to the time series information of user personality vector and sentence vector;

[0079] The personalized dialogue generation mode is used to input the word vector of the new input text into the personalized dialogue model to generate a personalized dialogue reply.

[0080] The decoding formula of the personalized dialogue model in the model generation mode is:

[0081] the s t = GRU decoder (s t-1...

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 relates to a personalized dialogue generation method and system based on user dialogue history. The personalized dialogue generation method comprises the following steps: S1, expressinga text of the user dialogue history as a sentence vector; S2, encoding the sentence vector to obtain a user personality vector, wherein the user personality vector comprises time sequence informationof the sentence vector; S3, generating a personalized dialogue model according to the time sequence information of the user personalized vector and the sentence vector; and S4, inputting the word vector of the new input text into the personalized dialogue model to generate a reply of the personalized dialogue. User modeling is carried out by using dialogue history, a user personality vector is calculated through user dialogue history reply, a user ID is prevented from being used for obtaining and updating the personality vector, and when a new user joins in, the dialogue history can be directly expressed as the personality vector.

Description

technical field [0001] The invention relates to a method and system for generating a personalized dialog based on user dialog history, and belongs to the technical field of artificial intelligence. Background technique [0002] With the development of end-to-end dialogue systems driven by data, personalized dialogue systems began to emerge. The chat-type personalized dialogue system is to give appropriate responses to the input in the open field, and at the same time, the output result can have the person's role information. In the application, whether it is a virtual social robot or an intelligent personal agent reply, it is required that the reply given by the dialogue model has character information, so it is necessary to give the model character setting (referred to as person setting in this article) to maintain the unity of character role information . At present, the methods for constructing persona for personalized dialogue systems are mainly divided into two types:...

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): G06F16/33G06F16/332G06F40/30G06K9/62
CPCG06F16/3329G06F40/30G06F16/3344G06F18/214Y02D10/00
Inventor 窦志成文继荣
Owner RENMIN UNIVERSITY OF CHINA
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com