Knowledge and data dual-driven streaming multi-mode graph learning method
The invention discloses a knowledge and data dual-driven streaming multi-modal graph learning method, and relates to the field of multi-modal graph learning, and the method comprises the following steps: training a lifelong learning knowledge pre-training network through collected multi-modal data;...
Saved in:
Main Authors | , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
12.03.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention discloses a knowledge and data dual-driven streaming multi-modal graph learning method, and relates to the field of multi-modal graph learning, and the method comprises the following steps: training a lifelong learning knowledge pre-training network through collected multi-modal data; carrying out decentration and normalization operation on the streaming multi-modal data by utilizing a data rectification module; complementing the missing data by using a data complementing module; a modal data identification module is used for identifying the complemented multi-modal data; inputting the multi-modal data in the data warehouse into representative prototype knowledge points of learning data in the trained lifelong learning knowledge pre-training network; and generating a bipartite graph according to the representative prototype knowledge points in the knowledge warehouse and the complemented multi-modal data by using a knowledge-driven graph generation module. According to the method, the problems t |
---|---|
Bibliography: | Application Number: CN202311758943 |