ML Anthology

Deterministic Annealing Variant of the EM Algorithm

Naonori Ueda, Ryohei Nakano
NeurIPS 1994 pp. 545-552
/neurips/1994/ueda1994neurips-deterministic/

Abstract

We present a deterministic annealing variant of the EM algorithm for maximum likelihood parameter estimation problems. In our approach, the EM process is reformulated as the problem of min(cid:173) imizing the thermodynamic free energy by using the principle of maximum entropy and statistical mechanics analogy. Unlike simu(cid:173) lated annealing approaches, this minimization is deterministically performed. Moreover, the derived algorithm, unlike the conven(cid:173) tional EM algorithm, can obtain better estimates free of the initial parameter values.

PDF NeurIPS Semantic Scholar

Cite

Text

Ueda and Nakano. "Deterministic Annealing Variant of the EM Algorithm." Neural Information Processing Systems, 1994.

Markdown

[Ueda and Nakano. "Deterministic Annealing Variant of the EM Algorithm." Neural Information Processing Systems, 1994.](https://mlanthology.org/neurips/1994/ueda1994neurips-deterministic/)

BibTeX

@inproceedings{ueda1994neurips-deterministic,
  title     = {{Deterministic Annealing Variant of the EM Algorithm}},
  author    = {Ueda, Naonori and Nakano, Ryohei},
  booktitle = {Neural Information Processing Systems},
  year      = {1994},
  pages     = {545-552},
  url       = {https://mlanthology.org/neurips/1994/ueda1994neurips-deterministic/}
}