Hypergraph Learning

Clustering, hyperedge prediction, and representation learning for higher-order graph structures.

Many real-world relationships are higher-order and are poorly captured by ordinary graphs. My doctoral research introduced principled notions of modularity for hypergraphs, scalable hypergraph clustering by iteratively reweighted modularity maximization, hyperedge prediction via resource allocation, and representation learning for multi-layer hypergraphs.

Representative publications (Kumar et al., 2019; Kumar et al., 2020; Kumar et al., 2020; Yadati et al., 2023).

References

2023

  1. LoG
    HEAL: Embedding Multi-layer Hypergraphs
    Naganand Yadati, Tarun Kumar, Deepak Maurya, and 2 more authors
    In Learning on Graphs Conference (LoG), 2023

2020

  1. Appl. Net. Sci.
    Hypergraph Clustering by Iteratively Reweighted Modularity Maximization
    Tarun Kumar, Sankaran Vaidyanathan, Harini Ananthapadmanabhan, and 2 more authors
    Applied Network Science, 2020
  2. WebSci
    HPRA: Hyperedge Prediction using Resource Allocation
    Tarun Kumar, K Darwin, Srinivasan Parthasarathy, and 1 more author
    In 12th ACM Conference on Web Science (WebSci), 2020

2019

  1. Complex Networks
    A New Measure of Modularity in Hypergraphs: Theoretical Insights and Implications for Effective Clustering
    Tarun Kumar, Sankaran Vaidyanathan, Harini Ananthapadmanabhan, and 2 more authors
    In International Conference on Complex Networks and Their Applications, 2019