Sum-Product Networks: A New Deep Architecture

Abstract

The key limiting factor in graphical model inference and learning is the complexity of the partition function. We thus ask the question: what are general conditions under which the partition function is tractable? The answer leads to a new kind of deep architecture, which we call sum-product networks (SPNs). SPNs are directed acyclic graphs with variables as leaves, sums and products as internal nodes, and weighted edges. We show that if an SPN is complete and consistent it represents the partition function and all marginals of some graphical model, and give semantics to its nodes. Essentially all tractable graphical models can be cast as SPNs, but SPNs are also strictly more general. We then propose learning algorithms for SPNs, based on backpropagation and EM. Experiments show that inference and learning with SPNs can be both faster and more accurate than with standard deep networks. For example, SPNs perform image completion better than state-of-the-art deep networks for this task. SPNs also have intriguing potential connections to the architecture of the cortex.

References in corpus (1)

• Learning Arithmetic Circuits

Cited by in corpus (24)

• Deep Learning in Neural Networks: An Overview
• MADE: Masked Autoencoder for Distribution Estimation
• DeepID-Net: multi-stage and deformable deep convolutional neural networks for object detection
• On the Relationship between Sum-Product Networks and Bayesian Networks
• On Relaxing Determinism in Arithmetic Circuits
• A Dynamic Programming Algorithm for Inference in Recursive Probabilistic Programs
• Accurate and Conservative Estimates of MRF Log-likelihood using Reverse Annealing
• SimNets: A Generalization of Convolutional Networks
• The Sum-Product Theorem: A Foundation for Learning Tractable Models
• The Tensor Memory Hypothesis
• Symmetry-Aware Marginal Density Estimation
• Complexity of Representation and Inference in Compositional Models with Part Sharing
• Optimisation of Overparametrized Sum-Product Networks
• Reified Context Models
• GSNs : Generative Stochastic Networks
• End-to-end learning potentials for structured attribute prediction
• Latent Dependency Forest Models
• Learning Tuple Probabilities
• The Complexity of Bayesian Networks Specified by Propositional and Relational Languages
• Coresets for Dependency Networks
• Weighted Positive Binary Decision Diagrams for Exact Probabilistic Inference
• Sum-Product Networks for Hybrid Domains
• Learning Tractable Probabilistic Models for Fault Localization
• Learning Large-Scale Topological Maps Using Sum-Product Networks