Inference and Learning in a Latent Variable Model for Beta Distributed Interval Data

Latent Variable Models (LVMs) are well established tools to accomplish a range of different data processing tasks. Applications exploit the ability of LVMs to identify latent data structure in order to improve data (e.g., through denoising) or to estimate the relation between latent causes and measurements in medical data. In the latter case, LVMs in the form of noisy-OR Bayes nets represent the standard approach to relate binary latents (which represent diseases) to binary observables (which represent symptoms). Bayes nets with binary representation for symptoms may be perceived as a coarse approximation, however. In practice, real disease symptoms can range from absent over mild and intermediate to very severe. Therefore, using diseases/symptoms relations as motivation, we here ask how standard noisy-OR Bayes nets can be generalized to incorporate continuous observables, e.g., variables that model symptom severity in an interval from healthy to pathological. This transition from binary to interval data poses a number of challenges including a transition from a Bernoulli to a Beta distribution to model symptom statistics. While noisy-OR-like approaches are constrained to model how causes determine the observables’ mean values, the use of Beta distributions additionally provides (and also requires) that the causes determine the observables’ variances. To meet the challenges emerging when generalizing from Bernoulli to Beta distributed observables, we investigate a novel LVM that uses a maximum non-linearity to model how the latents determine means and variances of the observables. Given the model and the goal of likelihood maximization, we then leverage recent theoretical results to derive an Expectation Maximization (EM) algorithm for the suggested LVM. We further show how variational EM can be used to efficiently scale the approach to large networks. Experimental results finally illustrate the efficacy of the proposed model using both synthetic and real data sets. Importantly, we show that the model produces reliable results in estimating causes using proofs of concepts and first tests based on real medical data and on images.

Probabilistic regression and Bayes nets

This chapter revises regression with the inclusion of uncertainty in the data in probabilistic models. It shows how modern probabilistic machine learning can be formulated. First, a simple stochastic generalization of the linear regression example is offered to introduce the formalism. This leads to the important maximum likelihood principle on which learning will be based. This concept is then generalized to non-linear problems in higher dimensions and the chapter relates this to Bayes nets. The chapter ends with a discussion about how such a probabilistic approach is related to deep learning.

Imprecise Bayesian Networks as Causal Models

This article considers the extent to which Bayesian networks with imprecise probabilities, which are used in statistics and computer science for predictive purposes, can be used to represent causal structure. It is argued that the adequacy conditions for causal representation in the precise context—the Causal Markov Condition and Minimality—do not readily translate into the imprecise context. Crucial to this argument is the fact that the independence relation between random variables can be understood in several different ways when the joint probability distribution over those variables is imprecise, none of which provides a compelling basis for the causal interpretation of imprecise Bayes nets. I conclude that there are serious limits to the use of imprecise Bayesian networks to represent causal structure.

Probabilistic language in indicative and counterfactual conditionals

This paper analyzes indicative and counterfactual conditionals that have in their consequents probability operators: probable, likely, more likely than not, 50% chance and so on. I provide and motivate a unified compositional semantics for both kinds of probabilistic conditionals using a Kratzerian syntax for conditionals and a representation of information based on Causal Bayes Nets. On this account, the only difference between probabilistic indicatives and counterfactuals lies in the distinction between conditioning and intervening. This proposal explains why causal (ir)relevance is crucial for probabilistic counterfactuals, and why it plays no direct role in probabilistic indicatives. I conclude with some complexities related to the treatment of backtracking counterfactuals and subtleties revealed by probabilistic language in the revision procedure used to create counterfactual scenarios. In particular, I argue that certain facts about the interaction between probability operators and counterfactuals motivate the use of Structural Equation Models (Pearl 2000) rather than the more general formalism of Causal Bayes Nets.

Analysis of evidence in international criminal trials using Bayesian Belief Networks

This article demonstrates how different actors in international criminal trials could utilise Bayesian Networks (‘Bayes Nets’), which are graphical models of the probabilistic relationships between hypotheses and pieces of evidence. We argue that Bayes Nets are potentially useful in both the examination of international criminal judgments and the processes of trial preparation and fact-finding before international criminal tribunals. With the use of a practical case study based on a completed case from the International Criminal Tribunal for the former Yugoslavia (ICTY), we illustrate how Bayes Nets could be used by international criminal tribunals to strengthen judges' confidence in their findings, to assist lawyers in preparing for trial, and to provide a tool for the assessment of international criminal tribunals' factual findings.