The rise of social networks has greatly contributed to creating information cascades. Over time, new nodes are added to the cascade network, which means the cascade network is dynamically variable. At the same time, there are often only a few nodes in the cascade network before new nodes join. Therefore, it becomes a key task to predict the diffusion after the dynamic cascade based on the small number of nodes observed in the previous period. However, existing methods are limited for dynamic short cascades and cannot combine temporal information with structural information well, so a new model, MetaCaFormer, based on meta-learning and the Transformer structure, is proposed in this paper for dynamic short cascade prediction. Considering the limited processing capability of traditional graph neural networks for temporal information, we propose a CaFormer model based on the Transformer structure, which inherits the powerful processing capability of Transformer for temporal information, while considering the neighboring nodes, edges and spatial importance of nodes, effectively combining temporal and structural information. At the same time, to improve the prediction ability for short cascades, we also fuse meta-learning so that it can be quickly adapted to short cascade data. In this paper, MetaCaFormer is applied to two publicly available datasets in different scenarios for experiments to demonstrate its effectiveness and generalization ability. The experimental results show that MetaCaFormer outperforms the currently available baseline methods.

Knowing how to effectively predict the scale of future information cascades based on the historical trajectory of information dissemination has become an important topic. It is significant for public opinion guidance; advertising; and hotspot recommendation. Deep learning technology has become a research hotspot in popularity prediction, but for complex social platform data, existing methods are challenging to utilize cascade information effectively. This paper proposes a novel end-to-end deep learning network CAC-G with cascade attention convolution (CAC). This model can stress the global information when learning node information and reducing errors caused by information loss. Moreover, a novel Dynamic routing-AT aggregation method is investigated and applied to aggregate node information to generate a representation of cascade snapshots. Then, the gated recurrent unit (GRU) is employed to learn temporal information. This study’s validity and generalization ability are verified in the experiments by applying CAC-G on two public datasets where CAC-G is better than the existing baseline methods.