Artificial Intelligence for Complex Network: Potential, Methodology and Application

Jingtao Ding,Chang Liu,Yu Zheng,Yunke Zhang,Zihan Yu,Ruikun Li,Hongyi Chen,Jinghua Piao,Huandong Wang,Jiazhen Liu,Yong Li

2024-02-23

Abstract:Complex networks pervade various real-world systems, from the natural environment to human societies. The essence of these networks is in their ability to transition and evolve from microscopic disorder-where network topology and node dynamics intertwine-to a macroscopic order characterized by certain collective behaviors. Over the past two decades, complex network science has significantly enhanced our understanding of the statistical mechanics, structures, and dynamics underlying real-world networks. Despite these advancements, there remain considerable challenges in exploring more realistic systems and enhancing practical applications. The emergence of artificial intelligence (AI) technologies, coupled with the abundance of diverse real-world network data, has heralded a new era in complex network science research. This survey aims to systematically address the potential advantages of AI in overcoming the lingering challenges of complex network research. It endeavors to summarize the pivotal research problems and provide an exhaustive review of the corresponding methodologies and applications. Through this comprehensive survey-the first of its kind on AI for complex networks-we expect to provide valuable insights that will drive further research and advancement in this interdisciplinary field.

Social and Information Networks,Artificial Intelligence,Machine Learning,Physics and Society

What problem does this paper attempt to address?

The paper aims to address several core challenges in the study of complex networks and explore the potential of artificial intelligence (AI) technologies in tackling these challenges. Specifically:

1. **Representation of Higher-Order Topological Properties**: Most current research on complex networks focuses on networks with pairwise connections, while interactions among three or more entities often exist in real systems. Therefore, studying how to represent these complex topological properties in higher-order structures has become a challenge.

2. **Unknown Dynamic Mechanisms**: Complex networks in the real world usually have highly complex and strongly coupled topological structures and individual dynamic characteristics. Traditional theoretical models, due to overly simplified assumptions, find it difficult to comprehensively describe their dynamic mechanisms, making it hard to accurately capture many important phenomena.

3. **Co-evolution Modeling of Node States and Interactions**: Due to the complex correlation between node states and their interactions, especially the correlation of node degrees, it becomes very difficult to theoretically model the co-evolution process of node states and edge relationships.

4. **Optimization of High-Dimensional Complex Networks**: The optimization problem on high-dimensional complex networks is a major challenge in the field of complex networks. Traditional methods lack analytical solutions when dealing with high-dimensional complex networks and have extremely high computational costs. Even approximate methods can lead to deviations from the theoretically optimal solution.

The paper systematically discusses how AI technologies (such as graph embedding, graph neural networks, etc.) can help overcome the above challenges and proposes a brand-new classification system for research problems, aiming to promote further development in this interdisciplinary field.