Empirical Inference Conference Paper 2009

Approximation Algorithms for Tensor Clustering

PDF Web
Thumb ticker sm thumb suvrit sra
Empirical Inference
no image
Empirical Inference

We present the first (to our knowledge) approximation algo- rithm for tensor clustering—a powerful generalization to basic 1D clustering. Tensors are increasingly common in modern applications dealing with complex heterogeneous data and clustering them is a fundamental tool for data analysis and pattern discovery. Akin to their 1D cousins, common tensor clustering formulations are NP-hard to optimize. But, unlike the 1D case no approximation algorithms seem to be known. We address this imbalance and build on recent co-clustering work to derive a tensor clustering algorithm with approximation guarantees, allowing metrics and divergences (e.g., Bregman) as objective functions. Therewith, we answer two open questions by Anagnostopoulos et al. (2008). Our analysis yields a constant approximation factor independent of data size; a worst-case example shows this factor to be tight for Euclidean co-clustering. However, empirically the approximation factor is observed to be conservative, so our method can also be used in practice.

Author(s): Jegelka, S. and Sra, S. and Banerjee, A.
Links:
Book Title: Algorithmic Learning Theory: 20th International Conference
Pages: 368-383
Year: 2009
Month: October
Day: 0
Editors: Gavalda, R. , G. Lugosi, T. Zeugmann, S. Zilles
Publisher: Springer
Bibtex Type: Conference Paper (inproceedings)
Address: Berlin, Germany
DOI: 10.1007/978-3-642-04414-4_30
Event Name: ALT 2009
Event Place: Porto, Portugal
Electronic Archiving: grant_archive
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik

BibTex

@inproceedings{5916,
  title = {Approximation Algorithms for Tensor Clustering},
  booktitle = {Algorithmic Learning Theory: 20th International Conference},
  abstract = {We present the first (to our knowledge) approximation algo-
  rithm for tensor clustering—a powerful generalization to basic 1D clustering. Tensors are increasingly common in modern applications dealing
  with complex heterogeneous data and clustering them is a fundamental
  tool for data analysis and pattern discovery. Akin to their 1D cousins,
  common tensor clustering formulations are NP-hard to optimize. But,
  unlike the 1D case no approximation algorithms seem to be known. We
  address this imbalance and build on recent co-clustering work to derive
  a tensor clustering algorithm with approximation guarantees, allowing
  metrics and divergences (e.g., Bregman) as objective functions. Therewith, we answer two open questions by Anagnostopoulos et al. (2008).
  Our analysis yields a constant approximation factor independent of data
  size; a worst-case example shows this factor to be tight for Euclidean
  co-clustering. However, empirically the approximation factor is observed
  to be conservative, so our method can also be used in practice.},
  pages = {368-383},
  editors = {Gavalda, R. , G. Lugosi, T. Zeugmann, S. Zilles},
  publisher = {Springer},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  address = {Berlin, Germany},
  month = oct,
  year = {2009},
  slug = {5916},
  author = {Jegelka, S. and Sra, S. and Banerjee, A.},
  month_numeric = {10}
}