Research Statement

In social networks, flow analytics identifies influential actors, maps information spread, and reveals communication bottlenecks. For technological networks like the internet and transportation systems, it maintains efficiency, robustness, and scalability. In biological networks, it offers insights protein interactions and disease spread. Whether in social, technological or biological networks, flow analytics provide vital capabilities by elucidating the pathways and dynamics of interconnected systems. This motivates developing advanced techniques to meet emerging needs in understanding and controlling real-world graph flows.

- Research Interests
  - Methodology
    - Higher-order Analysis over Graphs
    - Interaction between Network Flow and Structures
    - Spectral Theory on Graph Dynamics
  - Application
    - Physical Flow, e.g., Circuit, Traffic Dynamics
    - Virtual Flow, e.g., Social Message and Disease Spreading
    - Health: Brain, Genomics, and Epidemiology

Introduction: Graph machine learning, network science, and graph signal processing stand at the forefront of deciphering complex systems ranging from physical to virtual environments. The integration of physics and mathematics into AI models presents an innovative pathway to understanding and optimizing these systems. My research aims to pioneer methodologies and applications that bridge the gap between theoretical frameworks and practical solutions in various domains.

Research Interests:

  • Methodology:

    • Higher-order Analysis over Graphs: Leveraging advanced graph analysis techniques to uncover multi-level structural and dynamic patterns within networks, enhancing the comprehension and prediction of complex behaviors.
    • Interaction between Network Flow and Structures: Investigating the reciprocal effects between network flows (such as information, traffic, or fluid dynamics) and the underlying network structures to optimize flow efficiency and network resilience.
    • Spectral Theory on Graph Dynamics: Applying spectral theory to understand and manipulate graph dynamics, facilitating the analysis of temporal changes and the stability of networks.
    • Graph Bayesian Optimization: Employing Bayesian optimization methods in graph contexts to efficiently navigate the search space of network configurations for optimal performance in various tasks.
  • Application:

    • Physical Flow Applications: Addressing challenges in circuits and traffic dynamics to improve efficiency and reliability in physical infrastructures through the application of graph analytical methods.
    • Virtual Flow Applications: Enhancing the understanding and management of social message dissemination and disease spreading in virtual and real-world networks by applying novel graph-based frameworks.
    • Health Applications: Advancing the analysis and prediction in brain research, genomics, and epidemiology through the innovative use of graph machine learning techniques, aiming to uncover new insights into biological networks and disease propagation.

Objective: The overarching objective of my research is to develop and apply cutting-edge graph machine learning methodologies to dissect, understand, and optimize complex networks. By integrating principles from physics and mathematics, my work seeks to uncover fundamental insights and practical solutions that can significantly impact various domains, including but not limited to physical infrastructures, social systems, and healthcare.

Impact: The anticipated impact of this research includes the development of more efficient algorithms for network analysis, enhanced understanding of complex network behaviors, and the implementation of optimized strategies for managing network flows and dynamics. Ultimately, the integration of these methodologies into real-world applications promises significant advancements in system efficiency, societal wellbeing, and health outcomes.

Conclusion: In summary, my research endeavors to push the boundaries of graph machine learning, network science, and graph signal processing by developing innovative methodologies and applying them to critical applications. Through a rigorous, interdisciplinary approach, this work aims to provide novel insights and practical solutions for complex networked systems, contributing significantly to the advancement of science and technology.