Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and storage medium

A neural network and pre-training technology, applied in neural learning methods, biological neural network models, semantic analysis, etc., can solve the problem of no natural language reasoning and complete retrieval system processing in the retrieval of dialogues of different topics, so as to improve the retrieval effect. Effect

Active Publication Date: 2021-10-22
梁晨 +1
View PDF16 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0018] In order to solve the three major challenges of the existing technology for retrieval of conversations on different topics without natural language reasoning and without a complete retrieval system, the present invention proposes a pre-trained dual-attention neural network semantic inference dialog retrieval method, system, and retrieval equipment , storage medium, technical scheme of the present invention is as follows:

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
  • Pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and storage medium
  • Pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and storage medium
  • Pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0055] Embodiment 1: pre-training dual attention neural network semantic inference dialogue retrieval system, the system includes a data preprocessing module, a precoding module, a retrieval building module, a dialogue block module, a sorting module, an NLI training module and a model generation module;

[0056] The pre-processing module collects the dialogues and transmits them to the pre-encoding module for preliminary encoding. The dialogues are screened by the retrieval module and sent to the dialogue block module for differentiation, and the sorting module sorts the data and transmits them to the NLI training module. Network training, and finally a pre-trained dual attention neural network semantic inference dialogue retrieval system is established by the model generation module.

specific Embodiment approach 2

[0057] Specific implementation mode two: pre-training dual attention neural network semantic inference dialogue retrieval method, the method includes:

[0058] 1. Data preprocessing:

[0059] Firstly, multiple rounds of topic data are crawled on public anonymous online forums, and each of the chat content of all interlocutors is used as a node. The dialogues of the two identities of the host and the non-host always appear alternately. For each node, the nearest There is one and only one edge between the chat content of one or two groups of speakers with different identities, and its direction points to the speaking order. then as figure 1 Each path shown from the beginning to the end of a thread is a conversation derived from that forum thread.

[0060] Topics in online forums can be classified according to different tags, and data sets in movies, food, digital equipment, fashion and other fields can be fused separately. For user emotions, a similar treatment was carried ou...

specific Embodiment approach 3

[0150] Specific Embodiment 3: Those skilled in the art may use the methods mentioned in the above embodiments, and this embodiment may be provided as a method, system, or computer program product. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or a combination of software and hardware, and the modules may also be reorganized according to the computer logic structure. Furthermore, the present embodiments may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

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 pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and a storage medium, belongs to the field of man-machine language interaction, and aims to solve the problem that three challenges that no natural language inference exists for retrieval for different theme dialogues and no complete retrieval system exists for processing in the prior art. The system comprises a data preprocessing module, a pre-coding module, a retrieval establishing module, a dialogue partitioning module, a sorting module, an NLI training module and a model generating module. The preprocessing module records dialogues, the pre-coding module encodes the dialogues, the retrieval module screens the dialogues and then distinguishes and sorts the dialogues, the NLI training module performs neural network training by using a dual attention mechanism, and finally a model system is generated by a model; and under the same CPU, the processing speed is greatly improved, a large number of conversations can be processed in a short time, the accuracy is improved, and then reply sentences capable of well solving three challenges can be retrieved.

Description

technical field [0001] The invention relates to a pre-training dual attention neural network semantic inference dialogue retrieval method and system, retrieval equipment and a storage medium, belonging to the field of human-computer language interaction. Background technique [0002] In general, there are two types of dialogue systems: task-oriented dialogue and open-domain dialogue. Task-oriented dialog systems are designed for specific domains or tasks, such as flight reservations, hotel reservations, customer service and technical support, etc., and have been successfully applied in some practical applications. Building intelligent open-domain dialogue systems that can hold coherent and engaging conversations with humans has been a long-standing goal of artificial intelligence (AI). Early dialogue systems such as Eliza, Parry, and Alice, despite significantly improving machine intelligence, only worked well in restricted, stationary scenarios. The goal of open-domain di...

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): G06F16/332G06F16/35G06F40/30G06F40/295G06K9/62G06N3/04G06N3/08
CPCG06F16/3329G06F16/35G06F40/30G06F40/295G06N3/08G06N3/045G06F18/22G06F18/24Y02D10/00
Inventor 梁晨陈麒光耿健唐亚锋辛宇鑫
Owner 梁晨
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com