Online Algorithms for Sum-Product Networks with Continuous Variables

Priyank Jaini, Abdullah Rashwan, Han Zhao, Yue Liu, Ershad Banijamali, Zhitang Chen, Pascal Poupart
Proceedings of the Eighth International Conference on Probabilistic Graphical Models, PMLR 52:228-239, 2016.

Abstract

Sum-product networks (SPNs) have recently emerged as an attractive representation due to their dual interpretation as a special type of deep neural network with clear semantics and a tractable probabilistic graphical model. We explore online algorithms for parameter learning in SPNs with continuous variables. More specifically, we consider SPNs with Gaussian leaf distributions and show how to derive an online Bayesian moment matching algorithm to learn from streaming data. We compare the resulting generative models to stacked restricted Boltzmann machines and generative moment matching networks on real-world datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v52-jaini16, title = {Online Algorithms for Sum-Product Networks with Continuous Variables}, author = {Jaini, Priyank and Rashwan, Abdullah and Zhao, Han and Liu, Yue and Banijamali, Ershad and Chen, Zhitang and Poupart, Pascal}, booktitle = {Proceedings of the Eighth International Conference on Probabilistic Graphical Models}, pages = {228--239}, year = {2016}, editor = {Antonucci, Alessandro and Corani, Giorgio and Campos}, Cassio Polpo}, volume = {52}, series = {Proceedings of Machine Learning Research}, address = {Lugano, Switzerland}, month = {06--09 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v52/jaini16.pdf}, url = {https://proceedings.mlr.press/v52/jaini16.html}, abstract = {Sum-product networks (SPNs) have recently emerged as an attractive representation due to their dual interpretation as a special type of deep neural network with clear semantics and a tractable probabilistic graphical model. We explore online algorithms for parameter learning in SPNs with continuous variables. More specifically, we consider SPNs with Gaussian leaf distributions and show how to derive an online Bayesian moment matching algorithm to learn from streaming data. We compare the resulting generative models to stacked restricted Boltzmann machines and generative moment matching networks on real-world datasets.} }
Endnote
%0 Conference Paper %T Online Algorithms for Sum-Product Networks with Continuous Variables %A Priyank Jaini %A Abdullah Rashwan %A Han Zhao %A Yue Liu %A Ershad Banijamali %A Zhitang Chen %A Pascal Poupart %B Proceedings of the Eighth International Conference on Probabilistic Graphical Models %C Proceedings of Machine Learning Research %D 2016 %E Alessandro Antonucci %E Giorgio Corani %E Cassio Polpo Campos} %F pmlr-v52-jaini16 %I PMLR %P 228--239 %U https://proceedings.mlr.press/v52/jaini16.html %V 52 %X Sum-product networks (SPNs) have recently emerged as an attractive representation due to their dual interpretation as a special type of deep neural network with clear semantics and a tractable probabilistic graphical model. We explore online algorithms for parameter learning in SPNs with continuous variables. More specifically, we consider SPNs with Gaussian leaf distributions and show how to derive an online Bayesian moment matching algorithm to learn from streaming data. We compare the resulting generative models to stacked restricted Boltzmann machines and generative moment matching networks on real-world datasets.
RIS
TY - CPAPER TI - Online Algorithms for Sum-Product Networks with Continuous Variables AU - Priyank Jaini AU - Abdullah Rashwan AU - Han Zhao AU - Yue Liu AU - Ershad Banijamali AU - Zhitang Chen AU - Pascal Poupart BT - Proceedings of the Eighth International Conference on Probabilistic Graphical Models DA - 2016/08/15 ED - Alessandro Antonucci ED - Giorgio Corani ED - Cassio Polpo Campos} ID - pmlr-v52-jaini16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 52 SP - 228 EP - 239 L1 - http://proceedings.mlr.press/v52/jaini16.pdf UR - https://proceedings.mlr.press/v52/jaini16.html AB - Sum-product networks (SPNs) have recently emerged as an attractive representation due to their dual interpretation as a special type of deep neural network with clear semantics and a tractable probabilistic graphical model. We explore online algorithms for parameter learning in SPNs with continuous variables. More specifically, we consider SPNs with Gaussian leaf distributions and show how to derive an online Bayesian moment matching algorithm to learn from streaming data. We compare the resulting generative models to stacked restricted Boltzmann machines and generative moment matching networks on real-world datasets. ER -
APA
Jaini, P., Rashwan, A., Zhao, H., Liu, Y., Banijamali, E., Chen, Z. & Poupart, P.. (2016). Online Algorithms for Sum-Product Networks with Continuous Variables. Proceedings of the Eighth International Conference on Probabilistic Graphical Models, in Proceedings of Machine Learning Research 52:228-239 Available from https://proceedings.mlr.press/v52/jaini16.html.

Related Material