Deep kernel processes are a recently introduced class of deep Bayesian models that have the flexibility of neural networks, but work entirely with Gram matrices. They operate by alternately sampling a Gram matrix from a distribution over positive semi-definite matrices, and applying a deterministic transformation. When the distribution is chosen to be Wishart, the model is called a deep Wishart process (DWP). This particular model is of interest because its prior is equivalent to a deep Gaussian process (DGP) prior, but at the same time it is invariant to rotational symmetries, leading to a simpler posterior distribution. Practical inference in the DWP was made possible in recent work ("A variational approximate posterior for the deep Wishart process" Ober and Aitchison 2021a) where the authors used a generalisation of the Bartlett decomposition of the Wishart distribution as the variational approximate posterior. However, predictive performance in that paper was less impressive than one might expect, with the DWP only beating a DGP on a few of the UCI datasets used for comparison. In this paper, we show that further generalising their distribution to allow linear combinations of rows and columns in the Bartlett decomposition results in better predictive performance, while incurring negligible additional computation cost.
相關內容
Processing 是一門開源編程語言和與之配套的集成開發環境（IDE）的名稱。Processing 在電子藝術和視覺設計社區被用來教授編程基礎，并運用于大量的新媒體和互動藝術作品中。
Gaussian Process Networks (GPNs) are a class of directed graphical models which employ Gaussian processes as priors for the conditional expectation of each variable given its parents in the network. The model allows describing continuous joint distributions in a compact but flexible manner with minimal parametric assumptions on the dependencies between variables. Bayesian structure learning of GPNs requires computing the posterior over graphs of the network and is computationally infeasible even in low dimensions. This work implements Monte Carlo and Markov Chain Monte Carlo methods to sample from the posterior distribution of network structures. As such, the approach follows the Bayesian paradigm, comparing models via their marginal likelihood and computing the posterior probability of the GPN features. Simulation studies show that our method outperforms state-of-the-art algorithms in recovering the graphical structure of the network and provides an accurate approximation of its posterior distribution.
The number of modes in a probability density function is representative of the model's complexity and can also be viewed as the number of existing subpopulations. Despite its relevance, little research has been devoted to its estimation. Focusing on the univariate setting, we propose a novel approach targeting prediction accuracy inspired by some overlooked aspects of the problem. We argue for the need for structure in the solutions, the subjective and uncertain nature of modes, and the convenience of a holistic view blending global and local density properties. Our method builds upon a combination of flexible kernel estimators and parsimonious compositional splines. Feature exploration, model selection and mode testing are implemented in the Bayesian inference paradigm, providing soft solutions and allowing to incorporate expert judgement in the process. The usefulness of our proposal is illustrated through a case study in sports analytics, showcasing multiple companion visualisation tools. A thorough simulation study demonstrates that traditional modality-driven approaches paradoxically struggle to provide accurate results. In this context, our method emerges as a top-tier alternative offering innovative solutions for analysts.
Governments and industries have widely adopted differential privacy as a measure to protect users' sensitive data, creating the need for new implementations of differentially private algorithms. In order to properly test and audit these algorithms, a suite of tools for testing the property of differential privacy is needed. In this work we expand this testing suite and introduce R\'enyiTester, an algorithm that can verify if a mechanism is R\'enyi differentially private. Our algorithm computes computes a lower bound of the R\'enyi divergence between the distributions of a mechanism on neighboring datasets, only requiring black-box access to samples from the audited mechanism. We test this approach on a variety of pure and R\'enyi differentially private mechanisms with diverse output spaces and show that R\'enyiTester detects bugs in mechanisms' implementations and design flaws. While detecting that a general mechanism is differentially private is known to be NP hard, we empirically show that tools like R\'enyiTester provide a way for researchers and engineers to decrease the risk of deploying mechanisms that expose users' privacy.
Comparing different age estimation methods poses a challenge due to the unreliability of published results, stemming from inconsistencies in the benchmarking process. Previous studies have reported continuous performance improvements over the past decade using specialized methods; however, our findings challenge these claims. We argue that, for age estimation tasks outside of the low-data regime, designing specialized methods is unnecessary, and the standard approach of utilizing cross-entropy loss is sufficient. This paper aims to address the benchmark shortcomings by evaluating state-of-the-art age estimation methods in a unified and comparable setting. We systematically analyze the impact of various factors, including facial alignment, facial coverage, image resolution, image representation, model architecture, and the amount of data on age estimation results. Surprisingly, these factors often exert a more significant influence than the choice of the age estimation method itself. We assess the generalization capability of each method by evaluating the cross-dataset performance for publicly available age estimation datasets. The results emphasize the importance of using consistent data preprocessing practices and establishing standardized benchmarks to ensure reliable and meaningful comparisons. The source code is available at //github.com/paplhjak/Facial-Age-Estimation-Benchmark.
We provide a rigorous analysis of training by variational inference (VI) of Bayesian neural networks in the two-layer and infinite-width case. We consider a regression problem with a regularized evidence lower bound (ELBO) which is decomposed into the expected log-likelihood of the data and the Kullback-Leibler (KL) divergence between the a priori distribution and the variational posterior. With an appropriate weighting of the KL, we prove a law of large numbers for three different training schemes: (i) the idealized case with exact estimation of a multiple Gaussian integral from the reparametrization trick, (ii) a minibatch scheme using Monte Carlo sampling, commonly known as Bayes by Backprop, and (iii) a new and computationally cheaper algorithm which we introduce as Minimal VI. An important result is that all methods converge to the same mean-field limit. Finally, we illustrate our results numerically and discuss the need for the derivation of a central limit theorem.
We consider Bayesian linear regression with sparsity-inducing prior and design efficient sampling algorithms leveraging posterior contraction properties. A quasi-likelihood with Gaussian spike-and-slab (that is favorable both statistically and computationally) is investigated and two algorithms based on Gibbs sampling and Stochastic Localization are analyzed, both under the same (quite natural) statistical assumptions that also enable valid inference on the sparse planted signal. The benefit of the Stochastic Localization sampler is particularly prominent for data matrix that is not well-designed.
Traditionally, the Bayesian optimal auction design problem has been considered either when the bidder values are i.i.d., or when each bidder is individually identifiable via her value distribution. The latter is a reasonable approach when the bidders can be classified into a few categories, but there are many instances where the classification of bidders is a continuum. For example, the classification of the bidders may be based on their annual income, their propensity to buy an item based on past behavior, or in the case of ad auctions, the click through rate of their ads. We introduce an alternate model that captures this aspect, where bidders are \emph{a priori} identical, but can be distinguished based (only) on some side information the auctioneer obtains at the time of the auction. We extend the sample complexity approach of Dhangwatnotai, Roughgarden, and Yan (2014) and Cole and Roughgarden (2014) to this model and obtain almost matching upper and lower bounds. As an aside, we obtain a revenue monotonicity lemma which may be of independent interest. We also show how to use Empirical Risk Minimization techniques to improve the sample complexity bound of Cole and Roughgarden (2014) for the non-identical but independent value distribution case.
For many statistical experiments, there exists a multitude of optimal designs. If we consider models with uncorrelated observations and adopt the approach of approximate experimental design, the set of all optimal designs typically forms a multivariate polytope. In this paper, we mathematically characterize the polytope of optimal designs. In particular, we show that its vertices correspond to the so-called minimal optimum designs. Consequently, we compute the vertices for several classical multifactor regression models of the first and the second degree. To this end, we use software tools based on rational arithmetic; therefore, the computed list is accurate and complete. The polytope of optimal experimental designs, and its vertices, can be applied in several ways. For instance, it can aid in constructing cost-efficient and efficient exact designs.
The theory of mixed finite element methods for solving different types of elliptic partial differential equations in saddle-point formulation is well established since many decades. However, this topic was mostly studied for variational formulations defined upon the same finite-element product spaces of both shape- and test-pairs of primal variable-multiplier. Whenever these two product spaces are different the saddle point problem is asymmetric. It turns out that the conditions to be satisfied by the finite elements product spaces stipulated in the few works on this case may be of limited use in practice. The purpose of this paper is to provide an in-depth analysis of the well-posedness and the uniform stability of asymmetric approximate saddle point problems, based on the theory of continuous linear operators on Hilbert spaces. Our approach leads to necessary and sufficient conditions for such properties to hold, expressed in a readily exploitable form with fine constants. In particular standard interpolation theory suffices to estimate the error of a conforming method.
This paper introduces a smooth method for (structured) sparsity in $\ell_q$ and $\ell_{p,q}$ regularized optimization problems. Optimization of these non-smooth and possibly non-convex problems typically relies on specialized procedures. In contrast, our general framework is compatible with prevalent first-order optimization methods like Stochastic Gradient Descent and accelerated variants without any required modifications. This is accomplished through a smooth optimization transfer, comprising an overparametrization of selected model parameters using Hadamard products and a change of penalties. In the overparametrized problem, smooth and convex $\ell_2$ regularization of the surrogate parameters induces non-smooth and non-convex $\ell_q$ or $\ell_{p,q}$ regularization in the original parametrization. We show that our approach yields not only matching global minima but also equivalent local minima. This is particularly useful in non-convex sparse regularization, where finding global minima is NP-hard and local minima are known to generalize well. We provide a comprehensive overview consolidating various literature strands on sparsity-inducing parametrizations and propose meaningful extensions to existing approaches. The feasibility of our approach is evaluated through numerical experiments, which demonstrate that its performance is on par with or surpasses commonly used implementations of convex and non-convex regularization methods.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191