Statistical Translation, Heat Kernels and Expected Distances

Joshua V. Dillon, Yi Mao, Guy Lebanon, Jian Zhang

UAI 2007 pp. 93-100

doi:10.5555/3020488.3020500 /uai/2007/dillon2007uai-statistical/

Abstract

High dimensional structured data such as text and images is often poorly understood and misrepresented in statistical modeling. The standard histogram representation suffers from high variance and performs poorly in general. We explore novel connections between statistical translation, heat kernels on manifolds and graphs, and expected distances. These connections provide a new framework for unsupervised metric learning for text documents. Experiments indicate that the resulting distances are generally superior to their more standard counterparts.

UAI Semantic Scholar

Cite

Text

Dillon et al. "Statistical Translation, Heat Kernels and Expected Distances." Conference on Uncertainty in Artificial Intelligence, 2007. doi:10.5555/3020488.3020500

Markdown

[Dillon et al. "Statistical Translation, Heat Kernels and Expected Distances." Conference on Uncertainty in Artificial Intelligence, 2007.](https://mlanthology.org/uai/2007/dillon2007uai-statistical/) doi:10.5555/3020488.3020500

BibTeX

@inproceedings{dillon2007uai-statistical,
  title     = {{Statistical Translation, Heat Kernels and Expected Distances}},
  author    = {Dillon, Joshua V. and Mao, Yi and Lebanon, Guy and Zhang, Jian},
  booktitle = {Conference on Uncertainty in Artificial Intelligence},
  year      = {2007},
  pages     = {93-100},
  doi       = {10.5555/3020488.3020500},
  url       = {https://mlanthology.org/uai/2007/dillon2007uai-statistical/}
}