The empirical success of diffusion models in generative modelling has motivated theoretical work, including quantitative error bounds and qualitative analyses that characterise the different phases of denoising. We bring these two areas together by studying the adaptivity of diffusion models to the structured geometry of multimodal high-dimensional data that consists of multiple clusters in ℝ^D, each with its own low-dimensional structure, and inter-cluster separation depending on D. We employ K-mixture Gaussian distributions as a canonical framework to capture this geometry and establish two theoretical results. First, we interpret denoising as a dynamical Bayesian classifier: the mixture score is a posterior-weighted average of cluster-wise scores, and we show that, with high probability, the posterior class probabilities concentrate on a single cluster once the signal-to-noise ratio reaches the scale Θ(log (KD)/D). Second, by separately analysing the denoising process in its mixing and cluster-commitment phases, we prove that the KL error bound depends linearly on the maximum intrinsic dimension of a cluster, up to a logarithmic factor, even when K grows polynomially with D. This improves on ambient-dimensional bounds and extends existing low-dimensional adaptivity analyses to multimodal distributions with heterogeneous, approximately low-rank covariances.

Training-free conditional diffusion provides a flexible alternative to task-specific conditional model training, but existing samplers often allocate computation inefficiently: independent guided trajectories can vary widely in quality, and additional function evaluations along a single trajectory may not recover from poor early decisions. We propose Tempered Guided Diffusion (TGD), an annealed sequential Monte Carlo framework for training-free conditional sampling with diffusion priors. TGD targets tempered posterior distributions over the clean signal, using noisy diffusion states only as auxiliary variables for proposing reconstructions and propagating particles. Particles are reweighted by incremental likelihood ratios, resampled, and propagated across noise levels, concentrating computation on trajectories plausible under both the prior and observation. Under idealized exact-reconstruction assumptions, full TGD yields a consistent particle approximation to the posterior as the number of particles grows. For expensive reconstruction tasks, Accelerated TGD (A-TGD) retains early particle exploration but prunes to a single high-likelihood trajectory partway through sampling. Experiments on a controlled two-dimensional inverse problem and image inverse problems show improved posterior approximation and favorable wall-clock speed-quality tradeoffs over independent multi-trajectory baselines.

Quantifying uncertainty in detected changepoints is an important problem. However it is challenging as the naive approach would use the data twice, first to detect the changes, and then to test them. This will bias the test, and can lead to anti-conservative p-values. One approach to avoid this is to use ideas from post-selection inference, which conditions on the information in the data used to choose which changes to test. As a result this produces valid p-values; that is, p-values that have a uniform distribution if there is no change. Currently such methods have been developed for detecting changes in mean only. This paper presents two approaches for constructing post-selection p-values for detecting changes in variance. These vary depending on the method use to detect the changes, but are general in terms of being applicable for a range of change-detection methods and a range of hypotheses that we may wish to test.

With the advent of the Internet of Things, it is increasingly common to have large networks of sensors, where each sensor may collect different types of data, has limited local computing resources and the ability to transmit data to a central cloud. Detecting events that trigger changes in sensor data properties is a key concern. However, minimizing sensor-to-cloud communication might be necessary due either to privacy constraints or limited battery resources. To detect changes within such a network, we introduce a new method, mixFOCuS, which can detect changes in mixed-type data, where data from different sensors follow different, possibly non-Gaussian, distributions. This methods builds on the FOCuS algorithms, which are recent improvements of the classic approach of Page (1954), extending the approach to streaming data setting for distributed sensor networks. Our method does not require assuming known pre- and post-change parameters, yet is still efficient in both computation and communication, and suitable for detecting changes in real-time. We show the trade-off between reduced transmission frequency and detection power. Simulation results indicate improved power for mixed-type data and better performance than the existing works on Gaussian data.

Changepoint detection methods have seen considerable development in recent years, with online algorithms capable of identifying structural changes in streaming data in near real time. However, the majority of existing methods are designed under the assumption of IID observations, rendering them susceptible to either more false positives or longer detection delays when applied to data exhibiting temporal dependence, a common feature of many real-world data streams. In this article, we extend the generalised likelihood-ratio (GLR) statistic to autoregressive processes of order p, and adapt the focus algorithm to develop a computationally efficient online change detector. The resulting AR(p)-focus algorithm achieves an average computational cost of 𝒪(log n) per iteration, making it suitable for high-frequency data streams. Through simulation studies, the proposed approach is seen achieving greater detection power than IID-based tests when the underlying data exhibit temporal correlation. We further illustrate the practical utility of AR(p)-focus through an application to a real-world telecommunications dataset.

In online clustering problems, there is often a large amount of uncertainty over possible cluster assignments that cannot be resolved until more data are observed. This difficulty is compounded when clusters follow complex distributions, as is the case with text data. Sequential Monte Carlo (SMC) methods give a natural way of representing and updating this uncertainty over time, but have prohibitive memory requirements for large-scale problems. We propose a novel SMC algorithm that decomposes clustering problems into approximately independent subproblems, allowing a more compact representation of the algorithm state. Our approach is motivated by the knowledge base construction problem, and we show that our method is able to accurately and efficiently solve clustering problems in this setting and others where traditional SMC struggles.

Traditional compartmental models capture population-level dynamics but fail to characterize individual-level risk. The computational cost of exact likelihood evaluation for partially observed individual-based models, however, grows exponentially with the population size, necessitating approximate inference. Existing sampling-based methods usually require multiple simulations of the individuals in the population and rely on bespoke proposal distributions or summary statistics. We propose a deterministic approach to approximating the likelihood using categorical distributions. The approximate likelihood is amenable to automatic differentiation so that parameters can be estimated by maximization or posterior sampling using standard software libraries such as TensorFlow with little user effort. We prove the consistency of the maximum approximate likelihood estimator. We empirically test our approach on several classes of individual-based models for epidemiology: different sets of disease states, individual-specific transition rates, spatial interactions, under-reporting and misreporting. We demonstrate ground truth recovery and comparable log-likelihood values at substantially reduced cost compared to competitor methods. Finally, we show the scalability and effectiveness of our approach with a real-world application on the 2001 UK Foot-and-Mouth outbreak, where the simplicity of the CAL allows us to include 162,775 farms. Supplementary materials for this article are available online, including a standardized description of the materials available for reproducing the work.

We present an R and Python package for fast online changepoint detection in univariate and multivariate data streams for a variety of models. The package implements the focus family of algorithms, which compute the Generalised Likelihood Ratio test for a single changepoint exactly and efficiently, with a per-iteration cost of approximately log(n)^d for a d-dimensional sequence, without introducing approximations. This is achieved by exploiting a connection between the location of the changepoint candidates and the geometry of the data. The package supports a broad range of models from the natural exponential family, including Gaussian, Poisson, Binomial, Exponential and Gamma distributions, as well as a non-parametric detector based on the empirical cumulative distribution function and a detector for autoregressive data.

Understanding historical datasets, such as the England and Wales infant mortality data, for local government districts can provide valuable insights into our changing society. Such analyses can prove challenging in practice, due to frequent changes in the boundaries of local government districts for which records are collected. One solution adopted in the literature to overcome such practical challenges is to pre-process data using areal interpolation to render the units consistent over the time period of focus. However, such methods are prone to errors. In this paper we introduce a novel changepoint method to detect instances where interpolation performs poorly. We demonstrate the utility of our method on original data, and also demonstrate how correcting interpolation errors can affect the clustering of the infant mortality curves.

A state-space model is a time-series model that has an unobserved latent process from which we take noisy measurements over time. The observations are conditionally independent given the latent process and the latent process itself is Markovian. These properties lead to simplifications for the conditional distribution of the latent process given the parameters and the observations. This chapter looks at how we can leverage the properties of state-space models to construct efficient MCMC samplers. We consider a range of Gibbs-sampler schemes, including those which use the forward-backward algorithm to simulate from the full conditional of the latent process given the parameters. For models where the forward-backward algorithm is not applicable we look at particle MCMC algorithms that, given the parameters, use particle filters to approximately simulate from the latent process or estimate the likelihood of the observations. Throughout, we provide intuition and informally discuss theory about the properties of the model that impact the efficiency of the different algorithms and how approaches such as reparameterization can improve mixing.

State space models (SSMs) provide a flexible framework for modeling complex time series via a latent stochastic process. Inference for nonlinear, non-Gaussian SSMs is often tackled with particle methods that do not scale well to long time series. The challenge is two-fold: not only do computations scale linearly with time, as in the linear case, but particle filters additionally suffer from increasing particle degeneracy with longer series. Stochastic gradient MCMC methods have been developed to scale Bayesian inference for finite-state hidden Markov models and linear SSMs using buffered stochastic gradient estimates to account for temporal dependencies. We extend these stochastic gradient estimators to nonlinear SSMs using particle methods. We present error bounds that account for both buffering error and particle error in the case of nonlinear SSMs that are log-concave in the latent process. We evaluate our proposed particle buffered stochastic gradient using stochastic gradient MCMC for inference on both long sequential synthetic and minute-resolution financial returns data, demonstrating the importance of this class of methods.

Inference for high-dimensional hidden Markov models is challenging due to the exponential-in-dimension computational cost of calculating the likelihood. To address this issue, we introduce an innovative composite likelihood approach called “Simulation Based Composite Likelihood” (SimBa-CL). With SimBa-CL, we approximate the likelihood by the product of its marginals, which we estimate using Monte Carlo sampling. In a similar vein to approximate Bayesian computation (ABC), SimBa-CL requires multiple simulations from the model, but, in contrast to ABC, it provides a likelihood approximation that guides the optimization of the parameters. Leveraging automatic differentiation libraries, it is simple to calculate gradients and Hessians to not only speed up optimization but also to build approximate confidence sets. We present extensive empirical results which validate our theory and demonstrate its advantage over SMC, and apply SimBa-CL to real-world Aphtovirus data.

A graduate-level introduction to advanced topics in Markov chain Monte Carlo (MCMC), as applied broadly in the Bayesian computational context. The topics covered have emerged as recently as the last decade and include stochastic gradient MCMC, non-reversible MCMC, continuous time MCMC, and new techniques for convergence assessment. A particular focus is on cutting-edge methods that are scalable with respect to either the amount of data, or the data dimension, motivated by the emerging high-priority application areas in machine learning and AI. Examples are woven throughout the text to demonstrate how scalable Bayesian learning methods can be implemented. This text could form the basis for a course and is sure to be an invaluable resource for researchers in the field.

The increasing volume of data streams poses significant computational challenges for detecting changepoints online. Likelihood-based methods are effective, but their straightforward implementation becomes impractical online. We develop two online algorithms that exactly calculate the likelihood ratio test for a single changepoint in p-dimensional data streams by leveraging fascinating connections with computational geometry. Our first algorithm is straightforward and empirically quasi-linear. The second is more complex but provably quasi-linear: $\mathcal{O}(n\log(n)^{p+1})$ for $n$ data points. Through simulations, we illustrate, that they are fast and allow us to process millions of points within a matter of minutes up to $p=5$.

The rapid advancements of scalable methodologies have opened new avenues for analyzing complex spatio-temporal data, which is crucial in understanding dynamic environmental phenomena. This paper introduces a likelihood-based methodology for detecting abrupt changes in time in spatio-temporal processes, a field where traditional time series methods fall short. Unlike recent approaches, we do not make the unrealistic assumption that data is independent across changepoints. Instead, we use a recently proposed family of covariance models that allows nonstationarity in time, and we propose a Markov approximation to reduce the computational burden of calculating likelihoods under this model. We apply our method to two years of daily wind speed data from various synoptic weather stations in Ireland, identifying a significant changepoint on July 24, 2021, which aligns with a major shift in weather patterns. This application not only demonstrates the method’s utility in handling spatio-temporal datasets but also showcases its potential in broader environmental and climatic studies, offering a scalable solution for analyzing changing patterns in spatial data over time.

The computational cost of exact likelihood evaluation for partially observed and highly-heterogeneous individual-based models grows exponentially with the population size, therefore inference relies on approximations. Sampling-based approaches to this problem such as Sequential Monte Carlo or Approximate Bayesian Computation usually require simulation of every individual in the population multiple times and are heavily reliant on the design of bespoke proposal distributions or summary statistics, and can still scale poorly with population size. To overcome this, we propose a deterministic recursive approach to approximating the likelihood function using categorical distributions. The resulting algorithm has a computational cost as low as linear in the population size and is amenable to automatic differentiation, leading to simple algorithms for maximizing this approximate likelihood or sampling from posterior distributions. We prove consistency of the maximum approximate likelihood estimator of model parameters. We empirically test our approach on a range of models with various flavors of heterogeneity: different sets of disease states, individual-specific susceptibility and infectivity, spatial interaction mechanisms, under-reporting and mis-reporting. We demonstrate strong calibration performance, in terms of log-likelihood variance and ground truth recovery, and computational advantages over competitor methods. We conclude by illustrating the effectiveness of our approach in a real-world large-scale application using Foot-and-Mouth data from the 2001 outbreak in the United Kingdom.

We congratulate the authors on an important contribution to the exciting area of generative models: providing a simple framework for applying the idea of diffusion generative models to data defined on general spaces and for general noising processes.Diffusion models have shown remarkable empirical performance for several applications, most notably as a way of defining implicit models for images.However it is unclear, to us at least, why these methods are so successful.They typically involve fitting highly-parameterised models for the score, and it is unclear why one would expect to be able to find good parameter estimates for such models.Furthermore, such models may be sufficiently flexible to learn the exact denoising process for the samples they are trained on-thereby collapsing the diffusion model to a categorical sampler from the training data.Could the authors provide insight into why these methods work?Is there intuition as to what types of data their denoising Markov models will work well for?Diffusion models are often used when the data are supported on a low-dimensional manifold of the state-space.However, the KL divergence which these models implicitly minimize can be infinite when the supports of the distributions are disjoint.Can the proposed framework extend to objectives which do not have this issue?Or is the use of KL beneficial, in that it actually helps force the model to sample from close to the manifold?The proposed framework parameterises the generative process in terms of a function β(x; t), whose optimal value is the marginal density of the noised data at time t.However, in practice the authors use different parameterisations, mainly based on differences in log β(x; t).Can the authors give more intuition as to why this is the best choice?And do they think this would be the case in general?Finally, both parameterisations of the continuous-time Markov chain model revolve around log r(x, y; t) = log β(x; t) -log β(y; t).As far as we can tell, the authors do not 1

Recent work has suggested using Monte Carlo methods based on piecewise deterministic Markov processes (PDMPs) to sample from target distributions of interest. PDMPs are non-reversible continuous-time processes endowed with momentum, and hence can mix better than standard reversible MCMC samplers. Furthermore, they can incorporate exact sub-sampling schemes which only require access to a single (randomly selected) data point at each iteration, yet without introducing bias to the algorithm's stationary distribution. However, the range of models for which PDMPs can be used, particularly with sub-sampling, is limited. We propose approximate simulation of PDMPs with sub-sampling for scalable sampling from posterior distributions. The approximation takes the form of an Euler approximation to the true PDMP dynamics, and involves using an estimate of the gradient of the log-posterior based on a data sub-sample. We thus call this class of algorithms stochastic-gradient PDMPs. Importantly, the trajectories of stochastic-gradient PDMPs are continuous and can leverage recent ideas for sampling from measures with continuous and atomic components. We show these methods are easy to implement, present results on their approximation error and demonstrate numerically that this class of algorithms has similar efficiency to, but is more robust than, stochastic gradient Langevin dynamics.

Online changepoint detection aims to detect anomalies and changes in real-time in high-frequency data streams, sometimes with limited available computational resources. This is an important task that is rooted in many real-world applications, including and not limited to cybersecurity, medicine and astrophysics. While fast and efficient online algorithms have been recently introduced, these rely on parametric assumptions which are often violated in practical applications. Motivated by data streams from the telecommunications sector, we build a flexible nonparametric approach to detect a change in the distribution of a sequence. Our procedure, NP-FOCuS, builds a sequential likelihood ratio test for a change in a set of points of the empirical cumulative density function of our data. This is achieved by keeping track of the number of observations above or below those points. Thanks to functional pruning ideas, NP-FOCuS has a computational cost that is log-linear in the number of observations and is suitable for high-frequency data streams. In terms of detection power, NP-FOCuS is seen to outperform current nonparametric online changepoint techniques in a variety of settings. We demonstrate the utility of the procedure on both simulated and real data.