Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits

Robert Peharz, Steven Lang, Antonio Vergari, Karl Stelzner, Alejandro Molina, Martin Trapp, Guy Van Den Broeck, Kristian Kersting, Zoubin Ghahramani
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:7563-7574, 2020.

Abstract

Probabilistic circuits (PCs) are a promising avenue for probabilistic modeling, as they permit a wide range of exact and efficient inference routines. Recent “deep-learning-style” implementations of PCs strive for a better scalability, but are still difficult to train on real-world data, due to their sparsely connected computational graphs. In this paper, we propose Einsum Networks (EiNets), a novel implementation design for PCs, improving prior art in several regards. At their core, EiNets combine a large number of arithmetic operations in a single monolithic einsum-operation, leading to speedups and memory savings of up to two orders of magnitude, in comparison to previous implementations. As an algorithmic contribution, we show that the implementation of Expectation-Maximization (EM) can be simplified for PCs, by leveraging automatic differentiation. Furthermore, we demonstrate that EiNets scale well to datasets which were previously out of reach, such as SVHN and CelebA, and that they can be used as faithful generative image models.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-peharz20a, title = {Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits}, author = {Peharz, Robert and Lang, Steven and Vergari, Antonio and Stelzner, Karl and Molina, Alejandro and Trapp, Martin and Van Den Broeck, Guy and Kersting, Kristian and Ghahramani, Zoubin}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {7563--7574}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/peharz20a/peharz20a.pdf}, url = {http://proceedings.mlr.press/v119/peharz20a.html}, abstract = {Probabilistic circuits (PCs) are a promising avenue for probabilistic modeling, as they permit a wide range of exact and efficient inference routines. Recent “deep-learning-style” implementations of PCs strive for a better scalability, but are still difficult to train on real-world data, due to their sparsely connected computational graphs. In this paper, we propose Einsum Networks (EiNets), a novel implementation design for PCs, improving prior art in several regards. At their core, EiNets combine a large number of arithmetic operations in a single monolithic einsum-operation, leading to speedups and memory savings of up to two orders of magnitude, in comparison to previous implementations. As an algorithmic contribution, we show that the implementation of Expectation-Maximization (EM) can be simplified for PCs, by leveraging automatic differentiation. Furthermore, we demonstrate that EiNets scale well to datasets which were previously out of reach, such as SVHN and CelebA, and that they can be used as faithful generative image models.} }
Endnote
%0 Conference Paper %T Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits %A Robert Peharz %A Steven Lang %A Antonio Vergari %A Karl Stelzner %A Alejandro Molina %A Martin Trapp %A Guy Van Den Broeck %A Kristian Kersting %A Zoubin Ghahramani %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-peharz20a %I PMLR %P 7563--7574 %U http://proceedings.mlr.press/v119/peharz20a.html %V 119 %X Probabilistic circuits (PCs) are a promising avenue for probabilistic modeling, as they permit a wide range of exact and efficient inference routines. Recent “deep-learning-style” implementations of PCs strive for a better scalability, but are still difficult to train on real-world data, due to their sparsely connected computational graphs. In this paper, we propose Einsum Networks (EiNets), a novel implementation design for PCs, improving prior art in several regards. At their core, EiNets combine a large number of arithmetic operations in a single monolithic einsum-operation, leading to speedups and memory savings of up to two orders of magnitude, in comparison to previous implementations. As an algorithmic contribution, we show that the implementation of Expectation-Maximization (EM) can be simplified for PCs, by leveraging automatic differentiation. Furthermore, we demonstrate that EiNets scale well to datasets which were previously out of reach, such as SVHN and CelebA, and that they can be used as faithful generative image models.
APA
Peharz, R., Lang, S., Vergari, A., Stelzner, K., Molina, A., Trapp, M., Van Den Broeck, G., Kersting, K. & Ghahramani, Z.. (2020). Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:7563-7574 Available from http://proceedings.mlr.press/v119/peharz20a.html.

Related Material