Agglomerative Information Bottleneck

Abstract

We introduce a novel distributional clustering algorithm that max(cid:173) imizes the mutual information per cluster between data and giv(cid:173) en categories. This algorithm can be considered as a bottom up hard version of the recently introduced "Information Bottleneck Method". The algorithm is compared with the top-down soft ver(cid:173) sion of the information bottleneck method and a relationship be(cid:173) tween the hard and soft results is established. We demonstrate the algorithm on the 20 Newsgroups data set. For a subset of two news(cid:173) groups we achieve compression by 3 orders of magnitudes loosing only 10% of the original mutual information.

Cite

Text

Slonim and Tishby. "Agglomerative Information Bottleneck." Neural Information Processing Systems, 1999.

Markdown

[Slonim and Tishby. "Agglomerative Information Bottleneck." Neural Information Processing Systems, 1999.](https://mlanthology.org/neurips/1999/slonim1999neurips-agglomerative/)

BibTeX

@inproceedings{slonim1999neurips-agglomerative,
  title     = {{Agglomerative Information Bottleneck}},
  author    = {Slonim, Noam and Tishby, Naftali},
  booktitle = {Neural Information Processing Systems},
  year      = {1999},
  pages     = {617-623},
  url       = {https://mlanthology.org/neurips/1999/slonim1999neurips-agglomerative/}
}