Antonio Vergari

Simplifying, Regularizing and Strengthening Sum-Product Network Structure Learning

The need for feasible inference in Probabilistic Graphical Models PGMs has lead to tractable models like Sum-Product Networks SPNs. Their highly expressive power and their ability to provide exact and tractable inference make them very attractive for several real world applications, from computer vision to NLP. Recently, great attention around SPNs has focused on structure learning, leading to different algorithms being able to learn both the network and its parameters from data. Here, we enhance one of the best structure learner, LearnSPN, aiming to improve both the structural quality of the learned networks and their achieved likelihoods. Our algorithmic variations are able to learn simpler, deeper and more robust networks. These results have been obtained by exploiting some insights in the building process done by LearnSPN, by hybridizing the network adopting tree-structured models as leaves, and by blending bagging estimations into mixture creation. We prove our claims by empirically evaluating the learned SPNs on several benchmark datasets against other competitive SPN and PGM structure learners.

Learning Accurate Cutset Networks by Exploiting Decomposability

The rising interest around tractable Probabilistic Graphical Models is due to the guarantees on inference feasibility they provide. Among them, Cutset Networks (CNets) have recently been introduced as models embedding Pearl’s cutset conditioning algorithm in the form of weighted probabilistic model trees with tree-structured models as leaves. Learning the structure of CNets has been tackled as a greedy search leveraging heuristics from decision tree learning. Even if efficient, the learned models are far from being accurate in terms of likelihood. Here, we exploit the decomposable score of CNets to learn their structure and parameters by directly maximizing the likelihood, including the BIC criterion and informative priors on smoothing parameters. In addition, we show how to create mixtures of CNets by adopting a well known bagging method from the discriminative framework as an effective and cheap alternative to the classical EM. We compare our algorithms against the original variants on a set of standard benchmarks for graphical model structure learning, empirically proving our claims.

Visualizing and understanding Sum-Product Networks

Sum-Product Networks (SPNs) are deep tractable probabilistic models by which several kinds of inference queries can be answered exactly and in a tractable time. They have been largely used as black box density estimators, assessed by comparing their likelihood scores on different tasks. In this paper we explore and exploit the inner representations learned by SPNs. By taking a closer look at the inner workings of SPNs, we aim to better understand what and how meaningful the representations they learn are, as in a classic Representation Learning framework. We firstly propose an interpretation of SPNs as Multi-Layer Perceptrons, we then devise several criteria to extract representations from SPNs and finally we empirically evaluate them in several (semi-)supervised tasks showing they are competitive against classical feature extractors like RBMs, DBNs and deep probabilistic autoencoders, like MADEs and VAEs.

Learning Bayesian Random Cutset Forests

In the Probabilistic Graphical Model (PGM) community there is an interest around tractable models, i.e., those that can guarantee exact inference even at the price of expressiveness. Structure learning algorithms are interesting tools to automatically infer both these architectures and their parameters from data. Even if the resulting models are efficient at inference time, learning them can be very slow in practice. Here we focus on Cutset Networks (CNets), a recently introduced tractable PGM representing weighted probabilistic model trees with tree-structured models as leaves. CNets have been shown to be easy to learn, and yet fairly accurate. We propose a learning algorithm that aims to improve their average test log-likelihood while preserving efficiency during learning by adopting a random forest approach. We combine more CNets, learned in a generative Bayesian framework, into a generative mixture model. A thorough empirical comparison on real word datasets, against the original learning algorithms extended to our ensembling approach, proves the validity of our approach.

Fast and Accurate Density Estimation with Extremely Randomized Cutset Networks

Cutset Networks (CNets) are density estimators leveraging context-specific independencies recently introduced to provide exact inference in polynomial time. Learning a CNet is done by firstly building a weighted probabilistic OR tree and then estimating tractable distributions as its leaves. Specifically, selecting an optimal OR split node requires cubic time in the number of the data features, and even approximate heuristics still scale in quadratic time. We introduce Extremely Randomized Cutset Networks (XCNets), CNets whose OR tree is learned by performing random conditioning. This simple yet surprisingly effective approach reduces the complexity of OR node selection to constant time. While the likelihood of an XCNet is slightly worse than an optimally learned CNet, ensembles of XCNets outperform state-of-the-art density estimators on a series of standard benchmark datasets, yet employing only a fraction of the time needed to learn the competitors. Code and data related to this chapter are available at: https://github.com/nicoladimauro/cnet.

Density Estimators for Positive-Unlabeled Learning

Positive-Unlabeled (PU) learning works by considering a set of positive samples, and a (usually larger) set of unlabeled ones. This challenging setting requires algorithms to cleverly exploit dependencies hidden in the unlabeled data in order to build models able to accurately discriminate between positive and negative samples. We propose to exploit probabilistic generative models to characterize the distribution of the positive samples, and to label as reliable negative samples those that are in the lowest density regions with respect to the positive ones. The overall framework is flexible enough to be applied to many domains by leveraging tools provided by years of research from the probabilistic generative model community. Results on several benchmark datasets show the performance and flexibility of the proposed approach.

Alternative Variable Splitting Methods to Learn Sum-Product Networks

Sum-Product Networks (SPNs) are recent deep probabilistic models providing exact and tractable inference. SPNs have been successfully employed as density estimators in several application domains. However, learning an SPN from high dimensional data still poses a challenge in terms of time complexity. This is due to the high cost of determining independencies among random variables (RVs) and sub-populations among samples, two operations that are repeated several times. Even one of the simplest greedy structure learner, LearnSPN, scales quadratically in the number of the variables to determine RVs independencies. In this work we investigate approximate but fast procedures to determine independencies among RVs whose complexity scales in sub-quadratic time. We propose two procedures: a random subspace approach and one that adopts entropy as a criterion to split RVs in linear time. Experimental results prove that LearnSPN equipped by our splitting procedures is able to reduce learning and/or inference times while preserving comparable inference accuracy.