Abstract:Cascade prediction aims to estimate the popularity of information diffusion in complex networks, which is beneficial to many applications from identifying viral marketing to fake news propagation in social media, estimating the scientific impact (citations) of a new publication, and so on. How to effectively predict cascade growth size has become a significant problem. Most previous methods based on deep learning have achieved remarkable results, while concentrating on mining structural and temporal features from diffusion networks and propagation paths. Whereas, the ignorance of spread dynamic information restricts the improvement of prediction performance. In this paper, we propose a novel framework called Physics-informed graph convolutional network (PiGCN) for cascade prediction, which combines explicit features (structural and temporal features) and propagation dynamic status in learning diffusion ability of cascades. Specifically, PiGCN is an end-to-end predictor, firstly splitting a given cascade into sub-cascade graph sequence and learning local structures of each sub-cascade via graph convolutional network (GCN), then adopting multi-layer perceptron (MLP) to predict the cascade growth size. Moreover, our dynamic neural network, combining PDE-like equations and a deep learning method, is designed to extract potential dynamics of cascade diffusion, which captures dynamic evolution rate both on structural and temporal changes. To evaluate the performance of our proposed PiGCN model, we have conducted extensive experiment on two well-known large-scale datasets from Sina Weibo and ArXIv subject listing HEP-PH to verify the effectiveness of our model. The results of our proposed model outperform the mainstream model, and show that dynamic features have great significance for cascade size prediction.

CasFlow: Exploring Hierarchical Structures and Propagation Uncertainty for Cascade Prediction

Variational Information Diffusion for Probabilistic Cascades Prediction

Information Diffusion Prediction via Recurrent Cascades Convolution

Mining Subcascade Features for Cascade Outbreak Prediction in Big Networks.

CasCIFF: A Cross-Domain Information Fusion Framework Tailored for Cascade Prediction in Social Networks

Full-scale Cascade Dynamics Prediction with a Local-First Approach.

Information Cascade Popularity Prediction Via Probabilistic Diffusion

Information Diffusion Prediction via Exploiting Cascade Relationship Diversity.

Information Cascade Prediction of complex networks based on Physics-informed Graph Convolutional Network

Continual Information Cascade Learning

Information Cascades Modeling Via Deep Multi-Task Learning

GRASS: Learning Spatial–Temporal Properties from Chainlike Cascade Data for Microscopic Diffusion Prediction

Hierarchical Attention Neural Network for Information Cascade Prediction

Information Cascades Prediction with Graph Attention

CasGCN: Predicting future cascade growth based on information diffusion graph

Deep Collaborative Embedding for information cascade prediction

CCasGNN: Collaborative Cascade Prediction Based on Graph Neural Networks

CasTformer: A novel cascade transformer towards predicting information diffusion

Cascade modeling with multihead self-attention

Hierarchical Information Enhancement Network for Cascade Prediction in Social Networks

HierCas: Hierarchical Temporal Graph Attention Networks for Popularity Prediction in Information Cascades