Boosting as a Product of Experts

Narayanan U. Edakunni, Gary Brown, Tim Kovacs
Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence, PMLR R9:225-232, 2011.

Abstract

In this paper, we derive a novel probabilistic model of boosting as a Product of Experts. We re-derive the boosting algorithm as a greedy incremental model selection procedure which ensures that addition of new experts to the ensemble does not decrease the likelihood of the data. These learning rules lead to a generic boosting algorithm - POE- Boost which turns out to be similar to the AdaBoost algorithm under certain assumptions on the expert probabilities. The paper then extends the POEBoost algorithm to POEBoost.CS which handles hypothesis that produce probabilistic predictions. This new algorithm is shown to have better generalization performance compared to other state of the art algorithms.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR9-edakunni11a, title = {Boosting as a Product of Experts}, author = {Edakunni, Narayanan U. and Brown, Gary and Kovacs, Tim}, booktitle = {Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence}, pages = {225--232}, year = {2011}, editor = {Cozman, Fabio and Pfeffer, Avi}, volume = {R9}, series = {Proceedings of Machine Learning Research}, month = {14--17 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r9/main/assets/edakunni11a/edakunni11a.pdf}, url = {https://proceedings.mlr.press/r9/edakunni11a.html}, abstract = {In this paper, we derive a novel probabilistic model of boosting as a Product of Experts. We re-derive the boosting algorithm as a greedy incremental model selection procedure which ensures that addition of new experts to the ensemble does not decrease the likelihood of the data. These learning rules lead to a generic boosting algorithm - POE- Boost which turns out to be similar to the AdaBoost algorithm under certain assumptions on the expert probabilities. The paper then extends the POEBoost algorithm to POEBoost.CS which handles hypothesis that produce probabilistic predictions. This new algorithm is shown to have better generalization performance compared to other state of the art algorithms.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Boosting as a Product of Experts %A Narayanan U. Edakunni %A Gary Brown %A Tim Kovacs %B Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2011 %E Fabio Cozman %E Avi Pfeffer %F pmlr-vR9-edakunni11a %I PMLR %P 225--232 %U https://proceedings.mlr.press/r9/edakunni11a.html %V R9 %X In this paper, we derive a novel probabilistic model of boosting as a Product of Experts. We re-derive the boosting algorithm as a greedy incremental model selection procedure which ensures that addition of new experts to the ensemble does not decrease the likelihood of the data. These learning rules lead to a generic boosting algorithm - POE- Boost which turns out to be similar to the AdaBoost algorithm under certain assumptions on the expert probabilities. The paper then extends the POEBoost algorithm to POEBoost.CS which handles hypothesis that produce probabilistic predictions. This new algorithm is shown to have better generalization performance compared to other state of the art algorithms. %Z Reissued by PMLR on 04 October 2026.
APA
Edakunni, N.U., Brown, G. & Kovacs, T.. (2011). Boosting as a Product of Experts. Proceedings of the 27th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R9:225-232 Available from https://proceedings.mlr.press/r9/edakunni11a.html. Reissued by PMLR on 04 October 2026.

Related Material