Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model

2025-03-31

AI Summary

This work addresses the challenge of quantifying cross-platform propagation influence for short videos. We introduce XS-Video, the first large-scale, real-world cross-platform short-video propagation graph dataset—covering five major Chinese platforms, 535 topics, and 380K samples—and formally define the novel tasks of holistic interactive behavior modeling and cross-platform influence rating. Methodologically, we propose NetGPT, a large graph model that innovatively integrates heterogeneous graph neural networks with large language models (LLMs) via a three-stage curriculum training strategy: graph pretraining, LLM alignment fine-tuning, and multi-task refinement. Evaluated on XS-Video, NetGPT achieves state-of-the-art performance across both classification and regression metrics, significantly improving long-horizon propagation influence prediction accuracy. Our results empirically validate the effectiveness of jointly modeling structural graph topology and semantic content through deep LLM–graph integration.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Short-video platforms have gained immense popularity, captivating the interest of millions, if not billions, of users globally. Recently, researchers have highlighted the significance of analyzing the propagation of short-videos, which typically involves discovering commercial values, public opinions, user behaviors, etc. This paper proposes a new Short-video Propagation Influence Rating (SPIR) task and aims to promote SPIR from both the dataset and method perspectives. First, we propose a new Cross-platform Short-Video (XS-Video) dataset, which aims to provide a large-scale and real-world short-video propagation network across various platforms to facilitate the research on short-video propagation. Our XS-Video dataset includes 117,720 videos, 381,926 samples, and 535 topics across 5 biggest Chinese platforms, annotated with the propagation influence from level 0 to 9. To the best of our knowledge, this is the first large-scale short-video dataset that contains cross-platform data or provides all of the views, likes, shares, collects, fans, comments, and comment content. Second, we propose a Large Graph Model (LGM) named NetGPT, based on a novel three-stage training mechanism, to bridge heterogeneous graph-structured data with the powerful reasoning ability and knowledge of Large Language Models (LLMs). Our NetGPT can comprehend and analyze the short-video propagation graph, enabling it to predict the long-term propagation influence of short-videos. Comprehensive experimental results evaluated by both classification and regression metrics on our XS-Video dataset indicate the superiority of our method for SPIR.

Problem

• Proposes SPIR task for short-video influence rating
• Introduces XS-Video dataset for cross-platform propagation analysis
• Develops NetGPT model to predict video propagation influence

Innovation

• Cross-platform short-video dataset XS-Video
• Large Graph Model NetGPT for propagation analysis
• Three-stage training mechanism for graph-LLM integration