Boosting as a product of experts

Narayanan U. Edakunni, Gavin Brown, Tim Kovacs

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    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 pro- cedure which ensures that addition of new ex- perts 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 assump- tions on the expert probabilities. The pa- per then extends the POEBoost algorithm to POEBoost.CS which handles hypothesis that produce probabilistic predictions. This new algorithm is shown to have better generaliza- tion performance compared to other state of the art algorithms.
    Original languageEnglish
    Title of host publicationProceedings of the 27th Conference on Uncertainty in Artificial Intelligence, UAI 2011|Proc. Conf. Uncertainty Artif. Intell., UAI
    Pages187-194
    Number of pages7
    Publication statusPublished - 2011
    Event27th Conference on Uncertainty in Artificial Intelligence, UAI 2011 - Barcelona
    Duration: 1 Jul 2011 → …

    Conference

    Conference27th Conference on Uncertainty in Artificial Intelligence, UAI 2011
    CityBarcelona
    Period1/07/11 → …

    Fingerprint

    Dive into the research topics of 'Boosting as a product of experts'. Together they form a unique fingerprint.

    Cite this