Check patentability & draft patents in minutes with Patsnap Eureka AI!

Model training method, device, computer equipment, and computer-readable storage medium

A model training and model technology, applied in computing, neural learning methods, biological neural network models, etc., to achieve the effect of improving generalization ability

Active Publication Date: 2022-03-29
SICHUAN UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] One of the purposes of the present application is to provide a model training method, device, computer equipment and computer-readable storage medium to solve the problem of how to use target domain data to improve the generalization ability of the cross-domain slot filling model

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
  • Model training method, device, computer equipment, and computer-readable storage medium
  • Model training method, device, computer equipment, and computer-readable storage medium
  • Model training method, device, computer equipment, and computer-readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0054] Please refer to figure 1 , figure 1 It shows a schematic block diagram of the steps of a model training method provided by the embodiment of the present application.

[0055] Such as figure 1 As shown, the first model training method provided by the embodiment of the present application can be applied to a cross-domain slot filling model (Label-aware Transfer learning for Cross-domain Slot Filling, LTCS) incorporating label-aware transfer learning, including S110 to S140.

[0056] S110: Input a preset number of training samples into the embedded coding layer of the cross-domain slot filling model to obtain the hidden information of each word segment, wherein the training samples include samples in the first domain and samples in the second domain, and each training The samples all include real BIO tags.

[0057] In this embodiment, the BIO label marks each element as "B-X", "I-X" or "O". Among them, "B-X" indicates that the segment where this element is located belo...

Embodiment 2

[0097] Please refer to figure 2 , figure 2 A schematic structural block diagram of a model training device provided by an embodiment of the present application is shown. The model training device 500 includes an obtaining module 510 , a calculating module 520 and a training module 530 .

[0098] Wherein, the obtaining module 510 is configured to input a preset number of training samples into the embedded coding layer of the cross-domain slot filling model to obtain the hidden information of each word segment, wherein the training samples include the first domain samples and Second domain samples, each training sample includes real BIO labels;

[0099] The calculation module 520 is configured to calculate the maximum average difference value between the hidden information of the first domain samples and the second domain samples having the same real BIO label based on a first preset formula;

[0100] The calculation module 520 is further configured to add the maximum avera...

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 embodiment of the present application discloses a model training method, device, computer equipment, and computer-readable storage medium. The method is applied to a cross-domain slot-filling model that incorporates label-aware transfer learning, including: inputting a preset number of training samples into the embedded coding layer of the cross-domain slot-filling model to obtain hidden information for each word segment; based on the first A preset formula to calculate the maximum average difference value between the hidden information of the first domain sample and the second domain sample with the same real BIO label; add the maximum average difference values ​​corresponding to each real BIO label to obtain the maximum average The total value of the difference; with the goal of minimizing the maximum average total difference value, the cross-domain slot filling model is trained until the preset condition is met and the training is terminated. The model training method provided in this application uses target domain data to improve the generalization ability of the cross-domain slot filling model.

Description

technical field [0001] The present application relates to the field of intelligent voice technology, and in particular to a model training method, device, computer equipment and computer-readable storage medium. Background technique [0002] Spoken language understanding is an important part of natural language understanding, including domain classification, intent detection, and slot filling. Among them, the slot filling task is to extract the value of a well-defined attribute of a given entity from a large-scale corpus, that is, the slot filling task is used to identify task-related slot types in user utterances in a specific domain. [0003] The current cross-domain slot filling models use sufficient source domain data to achieve cross-domain slot filling, and do not make good use of the target domain data that is less marked. The generalization ability of the cross-domain slot filling model is weak. Therefore, how to use the target domain data to improve the generalizat...

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 Patents(China)
IPC IPC(8): G06F16/35G06F40/216G06F40/289G06N3/04G06N3/08
CPCG06F16/35G06F40/216G06F40/289G06N3/08G06N3/044G06N3/045
Inventor 周刚刘高硕琚生根
Owner SICHUAN UNIV
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