Greedy part-wise learning of sum-product networks

Robert Peharz, Bernhard Geiger, Franz Pernkopf

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

35 Citations (Scopus)


Sum-product networks allow to model complex variable interactions while still granting efficient inference. However, most learning algorithms proposed so far are explicitly or implicitly restricted to the image domain, either by assuming variable neighborhood or by assuming that dependent variables are related by their magnitudes over the training set. In this paper, we introduce a novel algorithm, learning the structure and parameters of sum-product networks in a greedy bottom-up manner. Our algorithm iteratively merges probabilistic models of small variable scope to larger and more complex models. These merges are guided by statistical dependence test, and parameters are learned using a maximum mutual information principle. In experiments our method competes well with the existing learning algorithms for sum-product networks on the task of reconstructing covered image regions, and outperforms these when neither neighborhood nor correlations by magnitude can be assumed.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases European Conference, ECML PKDD 2013, Prague, Czech Republic, September 23-27, 2013
Subtitle of host publicationProceedings Part 1
EditorsHendrik Blockeel, Kristian Kersting, Siegfried Nijssen, Filip Železný
ISBN (Electronic)978-3-642-40991-2
Publication statusPublished - 2013
Externally publishedYes
Event2013 European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2013) - Prague, Czech Republic
Duration: 23 Sep 201327 Sep 2013

Publication series

NameLecture Notes in Computer Science book series
PublisherSpringer Link
ISSN (Electronic)1973-2020
NameLecture Notes in Artificial Intelligence book sub series


Conference2013 European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2013)
Abbreviated titleECML PKDD 2013
Country/TerritoryCzech Republic


Dive into the research topics of 'Greedy part-wise learning of sum-product networks'. Together they form a unique fingerprint.

Cite this