Study different masking schemes for a universal marginaliser.
problem Understand how well a neural approximator learns conditional distributions.
method Compare networks trained with various masking schemes.
result Neural approximators perform differently based on the masking scheme.
UM trains a neural network to approximate marginal distributions in probabilistic programs.
problem High computational cost and lack of theoretical guarantees in inference methods for probabilistic programs.
method Combining samples from a probabilistic program prior with an augmentation method to train a neural network for any conditional marginal distribution.
result UM trains a single neural network to approximate any conditional marginal distribution, amortizing inference costs.
Bayesian causal inference method improves accuracy over traditional approaches.
problem Bayesian marginalisation over causal models is computationally infeasible.
method Decomposes structure marginalisation into causal orders and DAGs, using Gaussian processes for mechanisms and ARCO for orders.
result Method outperforms state-of-the-art in structure learning and inference.
Framework for uncertainty estimation in training parameters.
problem Estimating uncertainty in training parameters.
method Marginalizing hyperparameters as random variables, investigating various forms of marginalisation.
result Some marginalisations can reliably estimate uncertainty without extensive tuning.
Differentiable algorithm for time-series changepoint inference.
problem Posterior inference on time-series models with unknown changepoints.
method Differentiable algorithm based on dynamic programming.
result Improves time complexity from O(n^m) to O(mn).
Efficiently marginalizes over Gaussian Process kernels for better model flexibility and uncertainty.
problem Inefficient marginalization over Gaussian Process kernels for large datasets.
method Bayesian Quadrature scheme with maximum mean discrepancies and invariances between Spectral Mixture kernels.
result Achieves more accurate predictions and better calibrated uncertainty than state-of-the-art baselines.
Novel deep Gaussian process improves predictive uncertainty.
problem Flexible probabilistic data representations with tractable inference.
method Structured Gaussian variational family with marginalisation.
result Improved accuracy and calibrated uncertainty estimates.
Probabilistic graphical models are powerful tools which allow us to formalise our knowledge about the world and reason about its inherent uncertainty. There exist a considerable number of methods for performing inference in probabilistic graphical models; however, they can be computationally costly due to significant t…
We study a probabilistic numerical method for the solution of both boundary and initial value problems that returns a joint Gaussian process posterior over the solution. Such methods have concrete value in the statistics on Riemannian manifolds, where non-analytic ordinary differential equations are involved in virtual…
In this paper, we study classes of graphs with three types of edges that capture the modified independence structure of a directed acyclic graph (DAG) after marginalisation over unobserved variables and conditioning on selection variables using the m-separation criterion. These include MC, summary, and ancestral grap…
Bayesian Gaussian Processes improve exoplanet transit and Hubble constant inference.
problem Improving exoplanet transit and Hubble constant inference using Bayesian Gaussian Processes.
method Kernel-, mean- and noise-marginalised Gaussian Processes with evidence-based model comparison and transdimensional sampling.
result Inferred Hubble constant H0 values from cosmic chronometers, baryon acoustic oscillations and combined datasets are 66±6kms−1Mpc−1, 67±10kms−1Mpc−1 and 69±6kms−1Mpc−1, respectively. Bayesian neural networks benefit from fully marginalizing over all modes to improve generalization.
problem Bayesian neural networks suffer from multimodal posterior distributions that can lead to suboptimal generalization.
method Use appropriate Bayesian sampling tools to fully marginalize over all posterior modes.
result Training with full marginalization improves the ability of the network to reason between multiple candidate solutions.
Empowerment quantifies the influence an agent has on its environment. This is formally achieved by the maximum of the expected KL-divergence between the distribution of the successor state conditioned on a specific action and a distribution where the actions are marginalised out. This is a natural candidate for an intr…
This paper uses Nested Sampling to improve Gaussian Process uncertainty quantification.
problem Underestimating predictive uncertainty and overfitting in Gaussian Process models.
method Marginalises hyperparameters using Nested Sampling for spectral mixture kernels.
result Improves predictive performance and uncertainty quantification across various data sets.
Random Forests are reinterpreted as generative models to handle missing data and detect outliers.
problem Handling missing features and detecting outliers in Random Forests.
method Interpreting Random Forests as Probabilistic Circuits and applying marginalisation for missing data.
result GeDTs and GeFs can handle missing data and detect outliers under certain assumptions.
Integration over non-negative integrands is a central problem in machine learning (e.g. for model averaging, (hyper-)parameter marginalisation, and computing posterior predictive distributions). Bayesian Quadrature is a probabilistic numerical integration technique that performs promisingly when compared to traditional…
GNIs induce a regulariser that penalizes high-frequency components in neural network activations.
problem Understanding the regularizing effect of Gaussian noise injections on neural network activations.
method Deriving the explicit regularizer by marginalizing out injected noise and analyzing its effect in the Fourier domain.
result GNIs induce a regularizer that produces calibrated classifiers with large margins.
Single-pass method estimates neural network uncertainty.
problem Uncertainty estimation in deep learning requires multiple passes.
method Probabilistic reasoning over neural network depths.
result Single forward pass for uncertainty estimation.
Proposes a new Gaussian factor for probabilistic inference with degenerate settings.
problem Handling linear dependencies among random variables in Gaussian networks.
method Introduces a parametrised factor that relaxes the positive-definite constraint of the covariance matrix.
result Accurately accommodates degeneracies in probabilistic inference without significant computational overhead.
Marginalising out uncertain quantities within the internal representations or parameters of neural networks is of central importance for a wide range of learning techniques, such as empirical, variational or full Bayesian methods. We set out to generalise fast dropout (Wang & Manning, 2013) to cover a wider variety of …
Bayesian Quadrature improves ensembling for neural networks with dispersed likelihood peaks.
problem Ensembling neural networks struggles with dispersed, narrow peaks in likelihood surfaces.
method Uses Bayesian Quadrature to construct weighted ensembles of architectures.
result Empirically outperforms state-of-the-art baselines in test likelihood, accuracy, and expected calibration error.
A new Metropolis-Hastings algorithm uses Gaussian Processes to speed up sampling from complex models.
problem Sampling from computationally expensive probabilistic models.
method Two-stage Metropolis-Hastings algorithm with a Gaussian Process surrogate model.
result The approach learns the target distribution while sampling, eliminating the need for pre-training.
One-shot neural architecture search limits depth search space and prunes networks for better performance and uncertainty.
problem Finding optimal depth in residual networks for efficient training and inference.
method Formulated a variational objective to approximate the depth distribution and pruned networks based on this distribution.
result Pruned networks achieve competitive accuracy with unpruned networks and better uncertainty calibration.
Improved GP decoder training with SAS approximations.
problem Training expensive Gaussian process decoders is challenging and computationally expensive.
method Developed a new stochastic estimate of log-marginal likelihood based on cross-validation.
result SAS-GP improves robustness and reduces computational cost compared to variational autoencoders.
Complex systems can be modelled at various levels of detail. Ideally, causal models of the same system should be consistent with one another in the sense that they agree in their predictions of the effects of interventions. We formalise this notion of consistency in the case of Structural Equation Models (SEMs) by intr…
NDPs learn to sample from complex function distributions using neural networks and diffusion models.
problem Learning rich distributions over functions with neural networks.
method NDPs use denoising diffusion models and custom attention blocks to incorporate stochastic process properties.
result NDPs can capture functional distributions close to true Bayesian posteriors and outperform neural processes.
Model discrimination identifies a mathematical model that usefully explains and predicts a given system's behaviour. Researchers will often have several models, i.e. hypotheses, about an underlying system mechanism, but insufficient experimental data to discriminate between the models, i.e. discard inaccurate models. G…
We address the problem of causal discovery from data, making use of the recently proposed causal modeling framework of modular structural causal models (mSCM) to handle cycles, latent confounders and non-linearities. We introduce σ-connection graphs (σ-CG), a new class of mixed graphs (containing undirected, bidirected…
Pseudo-marginal Metropolis-Hastings (pmMH) is a powerful method for Bayesian inference in models where the posterior distribution is analytical intractable or computationally costly to evaluate directly. It operates by introducing additional auxiliary variables into the model and form an extended target distribution, w…
The issue of fairness in machine learning models has recently attracted a lot of attention as ensuring it will ensure continued confidence of the general public in the deployment of machine learning systems. We focus on mitigating the harm incurred by a biased machine learning system that offers better outputs (e.g. lo…
A new algorithm improves topic model accuracy in small data sets.
problem Limited data leads to inaccurate topic model identification.
method Developed a novel variational message passing algorithm (ALBU) for LDA.
result ALBU learns latent distributions more accurately than Variational Bayes (VB) in small data sets.
Statistical uncertainties are rarely incorporated in machine learning algorithms, especially for anomaly detection. Here we present the Bayesian Anomaly Detection And Classification (BADAC) formalism, which provides a unified statistical approach to classification and anomaly detection within a hierarchical Bayesian fr…
Transductive Adversarial Networks (TAN) is a novel domain-adaptation machine learning framework that is designed for learning a conditional probability distribution on unlabelled input data in a target domain, while also only having access to: (1) easily obtained labelled data from a related source domain, which may ha…
Novel approach to universal online learning for bounded losses, closing open problems.
problem Characterizing processes for universal online learning under non-i.i.d. conditions.
method Characterization of processes admitting strong and weak universal learning, introduction of optimistically universal learning rule.
result Introduction of a novel 1NN algorithm that is optimistically universal for bounded losses.
Universal Gaussian parity proven for 2D knots.
problem Proving universal Gaussian parity for 2D knots.
method Analyzing Gaussian parity on free 2D knots.
result Gaussian parity is universal for 2D knots.
Researchers found specific conformal groups for Einstein static universe models.
problem Understanding conformal groups of Einstein static universe models.
method Constructed explicit models for restricted conformal groups and universal covering groups.
result Determined all conformal Lorentz manifolds with maximal restricted conformal group dimension.
GCNNs gain rotation invariance with more training augmentation, making SVD-Universal more effective.
problem Improving robustness of GCNNs to adversarial attacks.
method SVD-Universal technique applied to GCNNs trained with larger rotations.
result SVD-Universal becomes more effective as GCNNs gain rotation invariance.
A novel VAE called πVAE models complex data with stochastic processes.
problem Efficient inference of stochastic processes for Bayesian deep learning with big data.
method πVAE is a finitely exchangeable, Kolmogorov consistent, continuous stochastic process VAE.
result πVAE achieves state-of-the-art performance in tasks like spatial interpolation.
The Gaussian Process Latent Variable Model (GP-LVM) is a non-linear probabilistic method of embedding a high dimensional dataset in terms low dimensional `latent' variables. In this paper we illustrate that maximum a posteriori (MAP) estimation of the latent variables and hyperparameters can be used for model selection…
The study of universal links in 3-manifolds and their properties.
problem Existence and characterization of universal links in 3-manifolds.
method Analyzing branched coverings and distinguishing between universal and complement universal links.
result Closed spherical 3-manifolds are the only ones admitting universal links.
Universal MLPs with a single hidden layer can learn any function.
problem Learning on various data structures like sequences, images, sets, and graphs.
method Using group theory, the paper proves the universality of a broad class of equivariant MLPs with a single hidden layer.
result Having a hidden layer on which the group acts regularly is sufficient for universal equivariance (invariance).
Constructs universal link invariants from intersections in configuration spaces.
problem Globalise topologically all coloured Jones polynomials and ADO polynomials.
method Defines new link invariants from graded intersections in configuration spaces.
result Recover all coloured Jones polynomials and ADO polynomials for links.
We construct universal Lefschetz fibrations, defined in analogy with classical universal bundles. We also introduce the cobordism groups of Lefschetz fibrations, and we see how these groups are quotients of the singular bordism groups via the universal Lefschetz fibrations.
We present a universal knot polynomials for 2- and 3-strand torus knots in adjoint representation, by universalization of appropriate Rosso-Jones formula. According to universality, these polynomials coincide with adjoined colored HOMFLY and Kauffman polynomials at SL and SO/Sp lines on Vogel's plane, and give their ex…
Revives Vogel's diagrammatic technique for universal Lie algebra computations.
problem The universality of Lie algebra quantities remains open, despite many being described.
method Diagrammatic algebra based on Vogel's Λ-algebra.
result Diagrammatic technique enables truly universal computations in Lie theory.
We prove Runge-type theorems and universality results for locally univalent holomorphic and meromorphic functions. Refining a result of M. Heins, we also show that there is a universal bounded locally univalent function on the unit disk. These results are used to prove that on any hyperbolic simply connected plane doma…
Simple technique turns any adversarial attack into a universal one using few test examples.
problem Creating universal adversarial attacks with minimal data.
method Universalization technique using few adversarial test examples and spectral properties.
result Simple universalization technique achieves comparable fooling rates to state-of-the-art methods.
Universal inequalities found for Laplacian eigenvalues on convex domains.
problem Finding bounds for Laplacian eigenvalues on convex domains.
method Established two universal inequalities.
result Found new bounds for Laplacian eigenvalues.