Knowledge question and answer dynamic graph construction method, knowledge question and answer method, storage medium, program product and electronic device

Abstract

The application provides a knowledge question answering dynamic graph construction method, a knowledge question answering method, a storage medium, a program product and an electronic device. The knowledge question answering dynamic graph construction method comprises the following steps: collecting multi-modal data containing at least time sequence image data to capture text semantics and locate positions; then performing semantic alignment through an adaptive gate fusion network of a multi-layer perception machine to generate a unified text vector; then using an improved multi-modal partition fusion network, performing cross fusion of entities, relationships and global features which are split from the unified text vector and visual features projected from the image to obtain multi-modal deep features; finally, inputting the enhanced deep features into a decoder to generate a timestamped multi-tuple set and inputting the timestamped multi-tuple set into a graph database to construct a knowledge question answering dynamic graph. Through multi-source data fusion and PFKAN knowledge graph construction, the application improves the semantic retrieval and reasoning robustness, and realizes dynamic adaptation and efficient decision-making in a multi-role scene.

Images