Title:
URBAN ENERGY MULTI-SOURCE DATA FUSION METHOD BASED ON DEEP LEARNING AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2024/082453
Kind Code:
A1
Abstract:
An urban energy multi-source data fusion method based on deep learning and a storage medium. Data fusion is performed on urban energy multi-source data mined by means of big data, and multi-scale multi-modal information fusion is performed by means of cross-modal Transformer, so as to achieve mutual cross-modal fusion of multi-source heterogeneous data types, and obtain fusion features for predicting the use amount of future energy and the required energy manufacturing amount. The present invention provides a multi-scale cooperative multi-modal transformer architecture, so as to improve the representation effect learned from a non-aligned multi-modal sequence; and allows the system to be more lightweight by enhancing the correlation degree in urban energy multi-source data fusion. The present invention can provide help for refined management of urban energy and scientific demand-side management, thus facilitating development of an intelligent decision function module of an urban energy big data platform.
Inventors:
ZHU ZHENGYI (CN)
XU HONGHUA (CN)
ZHANG WEIYA (CN)
LIANG LONG (CN)
MA JINJIE (CN)
ZHOU HENGJUN (CN)
WANG WENDI (CN)
QIAN XIN (CN)
YANG LINQING (CN)
XU HONGHUA (CN)
ZHANG WEIYA (CN)
LIANG LONG (CN)
MA JINJIE (CN)
ZHOU HENGJUN (CN)
WANG WENDI (CN)
QIAN XIN (CN)
YANG LINQING (CN)
Application Number:
PCT/CN2022/143923
Publication Date:
April 25, 2024
Filing Date:
December 30, 2022
Export Citation:
Assignee:
STATE GRID JIANGSU ELECTRIC POWER CO LTD NANJING POWER SUPPLY COMPANY (CN)
International Classes:
G06F18/25; G06N3/0464; G06Q10/04
Foreign References:
CN114463594A | 2022-05-10 | |||
CN114021628A | 2022-02-08 | |||
CN114239683A | 2022-03-25 | |||
US20220245424A1 | 2022-08-04 |
Other References:
LIANYANG MA; YU YAO; TAO LIANG; TONGLIANG LIU: "Multi-scale Cooperative Multimodal Transformers for Multimodal Sentiment Analysis in Videos", ARXIV.ORG, 16 June 2022 (2022-06-16), XP091252271
Attorney, Agent or Firm:
NANJING TIANYI PATENT AGENT CO., LTD. (CN)
Download PDF: