Abstract: We use the law of total variance to generate multiple expansions for the posterior predictive variance. These expansions are sums of terms involving conditional expectations and conditional variances and provide a quantification of the sources of predictive uncertainty. Since the posterior predictive variance is fixed given the model, it represents a constant quantity that is conserved over these expansions. The terms in the expansions can be assessed in absolute or relative sense to understand the main contributors to the length of prediction intervals. We quantify the term-wise uncertainty across expansions varying in the number of terms and the order of conditionates. In particular, given that a specific term in one expansion is small or zero, we identify the other terms in other expansions that must also be small or zero. We illustrate this approach to predictive model assessment in several well-known models. The Setting and Intuition Everyone uses prediction intervals (PI's) but few examine their structure or more precisely how they should be interpreted in the context of a model with multiple components. Often PI's seem overconfident (too narrow) or useless (too wide). Both frequentist and Bayesian practitioners routinely report PI's.

Neural Information Processing Systems http://nips.cc/

Concurrently, we introduce the associated optimisation algorithm, which is inspired by GECO [25]--the latter does not always lead to good encodings (e.g., Sec.

Learning concepts from natural high-dimensional data (e.g., images) holds potential

A.2 Proof of proposition 1 Let Pœ{B,D,E}, k be a valid kernel (assumptions of theorem 1) with K Inversion of conditional with Bayes rule gives: 'W œS As a complement, we now explicit the simple forms taken by the posterior limit graph in each case. A.3 Proof of theorem 2 We consider the following hierarchical model, for Nonetheless it can be simplified as we now show. We focus on finding the optimal eigenvectors first. Only the left term in (18) depends on R. The optimization problem for eigenvectors writes: min tr! Note that the identity permutation i.e. for i œ [n], (i) =i is optimal in this case as the ( We will choose this U in what follows as the sign of the axes do not influence the characterization of the final result in Z as a PCA embedding. Note that this solution is not unique if there are repeated eigenvalues.

Classical causal structure learning algorithms often assume no latent confounders. However, it is usually impossible to enumerate and measure all task-related variables in real-world scenarios. Neglecting latent confounders may lead to spurious correlations among observed variables.