A material vectorization method and system

By using the UIG-Embedding algorithm, combined with user behavior intent and material category system diagram representation, the diversity and adaptability issues of material vectorization methods in different business scenarios are solved, enabling the efficient application of material vectors in large enterprises.

CN116226502BActive Publication Date: 2026-06-19GUANGZHOU SHIYUAN ELECTRONICS CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGZHOU SHIYUAN ELECTRONICS CO LTD
Filing Date
2021-11-29
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing material vectorization methods fail to fully utilize material attributes, supply chain relationships, and user behavior information, resulting in insufficient diversity and adaptability in different business scenarios.

Method used

The fusion algorithm (UIG-Embedding) combines user behavior intent and material category system diagram representation. By acquiring user profiles, material profiles, and material category system structure data, it constructs a material diagram structure and trains a vectorized representation of materials by combining user behavior intent model and material click-through rate prediction model.

Benefits of technology

It achieves accurate and comprehensive representation of material vectors in different business scenarios, improves user retrieval efficiency and adaptability to the diversity of material information, and is suitable for multi-business and multi-scenario applications of large enterprises.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116226502B_ABST
    Figure CN116226502B_ABST
Patent Text Reader

Abstract

This application provides a material vectorization method and system. The method includes: acquiring user profile data, material profile data, user behavior data, and material category structure data; constructing behavioral intent sequences and material click sequences based on user behavior data, and constructing a material graph structure based on the material category structure data; pre-training a material graph representation model by combining the material graph structure and material profile data; training a user behavior intent model by combining user profile data and behavioral intent sequences, and training a material click-through rate prediction model by combining the user behavior intent model, material click sequences, material graph representation model, and material profile data; and fine-tuning the parameters of the material graph representation model through backpropagation; after the material graph representation model gradually converges, using the fine-tuned material vectors as the vectors of the entire material system. This application offers more accurate and comprehensive vector representation, more comprehensive sequence representation, better model interpretability, and greater versatility.
Need to check novelty before this filing date? Find Prior Art