Research

(Last updated in Feb. 2017)
My research can be summarized by the following five categories. The applications include data science, network analysis, signal processing, applied machine learning, data mining, cyber security, cyber attack and defense modeling, and event propagation and control in networks.

  1. Artificial Intelligence Tasks via Graph Data Analytics 
    • Reading comprehension (text mining)
    • Question and answer
  2. Fundamental Limits for Graph Data Analysis
    • Established fundamental principles governing the performance of spectral algorithms on community detection and graph clustering for single-layer and multilayer graphs
    • Proposed theory-driven automated model order selection methods (AMOS & MIMOSA) for clustering on graphs
  3. Network Optimization via Actions on Graphs
    • Developed an effective edge-rewiring method for enhancing topological network resilience
    • Proposed a new centrality measure, local Fiedler vector centrality (LFVC), for connectivity assessment on nodes and edges
    • Proposed a framework on deep community detection and extraction for network data exploration
    • Developed effective methods to enhance cyber security against explicit and lateral cyber attacks
    • Established performance guarantees of greedy algorithms on graphs
    • Developed a theoretical framework for accelerating distributed optimization over networks
  4. Event Propagation and Control in Networks
    • Proposed a novel rate-reliability-delay adaptation scheme via network coding for multi-path routing in communication networks
    • Developed analytical models for information dissemination dynamics in heterogeneous networks (network of networks, interconnected systems)¬†
    • Proposed effective methods for information propagation control, particularly for malware propagation control and epidemic routing
    • Proposed an effective method for identifying influential links for event propagation in online social networks
  5. Applications to Data Science
    • Proposed FEAST, an automated feature selection framework for compilation tasks in computer programs
    • Developed a framework for integrating mobile sensing with crowdsourcing algorithms
    • Developed an efficient incremental computation method (Incremental-IO) for spectral clustering of increasing orders