Optimisation of English learning paths in multimodal social networks using graph neural networks

Personalised English learning based on multimodal data is in increasing demand given the fast development of information technology and social media.It is challenging to fully understand learners' behavioural traits and learning demands in multimodal social environments, nevertheless, since conventional approaches of path optimisation usually concentrate on one modality or neglect social ties.This work presents a multimodal fusion path optimisation model based on graph neural network (GNN), i.e., MMP-GENLPO, to handle the above challenges.By combining multimodal elements and applying path optimisation techniques, MMP-GENLPO efficiently increases the structural reasonableness and personalised adaptive capacity of learning paths.MMP-GENLPO has good application prospects and promotion value and verifies its efficacy and practicality in intelligent English learning recommendation by outperforming several comparative models in four key metrics by means of experimental validation on real multimodal social learning datasets.