多模态大模型驱动的学科知识图谱进化及教育应用

Evolution and E ducational A pplication of Discipline Knowledge Graph Driven by Multimodal Large Model

当前,以GPT-4为代表的多模态大模型正在促进通用人工智能向多模态融合的方向发展。受益于多模态思维链、涌现能力和提示工程等应用技术,学科知识图谱能够更好地形成一种将抽象的符号概念和多模态的数据加以关联的智能扩展范式,有效解决现有图谱构建中存在的诸多不足,进一步推动多模态学科知识图谱的创生,为智慧教育服务的发展提供有力的支持。基于此,文章首先解构多模态大模型与学科知识图谱的关系;然后,文章探析了多模态大模型对学科知识图谱的驱动前提,从进化角度探讨多模态学科知识图谱的基本内涵和构建框架;最后,文章提出多模态学科知识图谱的四大教育应用场景,包括推动教育资源多元聚合、助力智能教学产品开发、赋能学科资源个性推荐以及促进人机协同智慧教学,以期为学科知识图谱的相关研究提供借鉴与启发。

At present,the multimodal large model represented by GPT-4 is promoting the development of general artificial intelligence in the direction of multimodal integration.Benefiting from application technologies such as multimodal thought chain,emergent abilities and prompt engineering,discipline knowledge map can better form an intelligent expansion paradigm that correlate s abstract symbolic concepts with multimodal data,which can effectively solve many shortcomings in the construction of existing maps,and further promote the creation of multimodal discipline knowledge graph s,and provide strong support for the development of smart education services.Based on this,this paper firstly deconstructed the relationship between multimodal large model and discipline knowledge graph.Then,this paper explored the driving premise of multimodal large model on discipline knowledge graph,and discussed the basic connotation and construction framework of multimodal discipline knowledge graph from the evolutionary perspective.Finally,four educational application scenarios of multimodal discipline knowledge graph were proposed,including promoting the diverse aggregation of educational resources,helping the development of intelligent teaching products,empowering the personalized recommendation of discipline resources,and promoting the smart teaching of human-machine collaboration,expecting to provide reference and enlightenment for the related research of discipline knowledge graph.