Directed Graph Embedding: An Algorithm Based on Continuous Limits of Laplacian-Type Operators

Abstract

This paper considers the problem of embedding directed graphs in Euclidean space while retaining directional information. We model the observed graph as a sample from a manifold endowed with a vector field, and we design an algo- rithm that separates and recovers the features of this process: the geometry of the manifold, the data density and the vector field. The algorithm is motivated by our analysis of Laplacian-type operators and their continuous limit as generators of diffusions on a manifold. We illustrate the recovery algorithm on both artificially constructed and real data.

Cite

Text

Perrault-joncas and Meila. "Directed Graph Embedding: An Algorithm Based on Continuous Limits of Laplacian-Type Operators." Neural Information Processing Systems, 2011.

Markdown

[Perrault-joncas and Meila. "Directed Graph Embedding: An Algorithm Based on Continuous Limits of Laplacian-Type Operators." Neural Information Processing Systems, 2011.](https://mlanthology.org/neurips/2011/perraultjoncas2011neurips-directed/)

BibTeX

@inproceedings{perraultjoncas2011neurips-directed,
  title     = {{Directed Graph Embedding: An Algorithm Based on Continuous Limits of Laplacian-Type Operators}},
  author    = {Perrault-joncas, Dominique C. and Meila, Marina},
  booktitle = {Neural Information Processing Systems},
  year      = {2011},
  pages     = {990-998},
  url       = {https://mlanthology.org/neurips/2011/perraultjoncas2011neurips-directed/}
}