Importance Sampling Research Articles

RPEM: Randomized Monte Carlo parametric expectation maximization algorithm.

Inspired from quantum Monte Carlo, by sampling discrete and continuous variables at the same time using the Metropolis-Hastings algorithm, we present a novel, fast, and accurate high performance Monte Carlo Parametric Expectation Maximization (MCPEM) algorithm. We named it Randomized Parametric Expectation Maximization (RPEM). We compared RPEM with NONMEM's Importance Sampling Method (IMP), Monolix's Stochastic Approximation Expectation Maximization (SAEM), and Certara's Quasi-Random Parametric Expectation Maximization (QRPEM) for a realistic two-compartment voriconazole model with ordinary differential equations using simulated data. We show that RPEM is as fast and as accurate as the algorithms IMP, QRPEM, and SAEM for the voriconazole model in reconstructing the population parameters, for the normal and log-normal cases.

A multi-modal importance sampling approach for the probabilistic safety assessment of automated driver assistance systems

Abstract In this paper, we present a stochastic approach for the reliability evaluation of specific traffic scenarios as one component in the validation procedure of Advanced Driver Assistance Systems (ADAS). In this analysis, the control device is represented as a simulation model using software-in-the-loop technology. Specific inputs of this simulated controller are modeled as scalar random inputs. Based on a definition of a failure criterion, the well known reliability method can be applied. In the present paper, a variance reduced importance sampling strategy for multiple failure regions is presented, which was developed for a scenario-based safety assessment framework.

A distributed particle filter with sampling-based consensus density fusion for speaker tracking in distributed microphone networks

This paper proposes a novel distributed particle filter with sampling-based consensus density fusion for speaker tracking in distributed microphone networks (DMNs). Firstly, to combat the reverberation and ambient noise, multi-hypothesis likelihood model is applied to describe the likelihood distribution in the state-space model of speaker tracking problem. Then, the sequential importance resampling (SIR) is performed on each node to obtain the local posteriors of the state vector of the speaker. Since the non-Gaussian noise and the reverberation may make local posteriors presenting non-Gaussian and multi-modal characters, in this paper, the local posterior densities are approximated by Gaussian mixtures whose parameters are further transmitted to neighboring nodes to complete the distributed fusion based on the consensus method. We investigate a novel adaptive multiple importance sampling (AMIS) algorithm combining parallel tempering Monte Carlo (PTMC) samplers to achieve the sampling-based consensus density fusion of the non-Gaussian local posteriors and estimate the global posterior locally on each node. At last, the optimal global estimation of the speaker state vector could be calculated by using the estimated global posterior on each node in DMNs. Simulation and real-world recording experiment results show that the proposed speaker tracking method exhibits satisfactory tracking performance under reverberant and noisy environments. Moreover, it is easy and feasible to extend the proposed tracking algorithm to solve the multi-speaker tracking problem.

Mathematical Modelling of Parasite Dynamics: A Stochastic Simulation-Based Approach and Parameter Estimation via Modified Sequential-Type Approximate Bayesian Computation

The development of mathematical models for studying newly emerging and re-emerging infectious diseases has gained momentum due to global events. The gyrodactylid-fish system, like many host-parasite systems, serves as a valuable resource for ecological, evolutionary, and epidemiological investigations owing to its ease of experimental manipulation and long-term monitoring. Although this system has an existing individual-based model, it falls short in capturing information about species-specific microhabitat preferences and other biological details for different Gyrodactylus strains across diverse fish populations. This current study introduces a new individual-based stochastic simulation model that uses a hybrid τ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ au $$\\end{document}-leaping algorithm to incorporate this essential data, enhancing our understanding of the complexity of the gyrodactylid-fish system. We compare the infection dynamics of three gyrodactylid strains across three host populations. A modified sequential-type approximate Bayesian computation (ABC) method, based on sequential Monte Carlo and sequential importance sampling, is developed. Additionally, we establish two penalised local-linear regression methods (based on L1 and L2 regularisations) for ABC post-processing analysis to fit our model using existing empirical data. With the support of experimental data and the fitted mathematical model, we address open biological questions for the first time and propose directions for future studies on the gyrodactylid-fish system. The adaptability of the mathematical model extends beyond the gyrodactylid-fish system to other host-parasite systems. Furthermore, the modified ABC methodologies provide efficient calibration for other multi-parameter models characterised by a large set of correlated or independent summary statistics.

Temperature Control of Fuel Cell Based on PEI-DDPG

Proton exchange membrane fuel cells (PEMFCs) constitute nonlinear systems that are challenging to model accurately. Therefore, a controller with robustness and adaptability is imperative for temperature control within the PEMFC stack. This paper introduces a data-driven controller utilizing deep reinforcement learning for stack temperature control. Given the PEMFC system’s characteristics, such as nonlinearity, uncertainty, and environmental conditions, we propose a novel deep reinforcement learning algorithm—the deep deterministic policy gradient with priority experience playback and importance sampling method (PEI-DDPG). Algorithm design incorporates technologies such as priority experience playback, importance sampling, and optimized sample data storage structure, enhancing the controller’s performance. Simulation results demonstrate the proposed algorithm’s superior effectiveness in temperature control for PEMFC, leveraging the PEI-DDPG algorithm’s high adaptability and robustness. The proposed algorithm’s effectiveness is additionally validated on the RT-LAB experimental platform. The proposed PEI-DDPG algorithm reduces the average adjustment time by 8.3%, 17.13%, and 24.56% and overshoots by 2.12 times, 4.16 times, and 4.32 times compared to the TD3, GA-PID, and PID algorithms, respectively.

Statistical inference of comparative generalized inverted exponential populations under joint adaptive progressive type-II censored samples

Design of efficient test plans to assessing product reliability rapidly, is a critical step to obtain product reliability accurately. In this paper, we are adopted the problem of statistical inference of two generalized inverted exponential populations in a competing duration. To save the balance between the optimal test time and the number of failures needing for statistical inference, joint adaptive type-II hybrid progressive censoring scheme is applied. The model parameters are estimated with classical methods (maximum likelihood and bootstrap) for point estimate and the corresponding confidence intervals. Also, Bayes estimators of model parameters are computed under importance sample technique with the corresponding Bayes credible intervals. A real data sets are analyzed for illustrating purposes. The results are compared and assessed through Monte Carlo simulation study.

System reliability analysis with small failure probability based on relevant vector machine and Meta-IS idea

Structural failure is a complex system problem and traditional methods face difficulties in solving system reliability problems with small failure probability. This paper proposes an efficient sampling simulation method termed SDMI, which integrates metamodel-based importance sampling (Meta-IS) with spherical decomposition (SD). SDMI transforms the failure probability into a product of spherical augmented failure probability and a correction factor. The implementation of SDMI for each component further facilitates the proposition of a system probability classification function. The integration of the relevant vector machine (RVM) leads to the proposal of a more efficient methodology, termed RVM-SDMI. The efficacy and precision of the proposed methods have been validated through four numerical examples and two engineering examples. The results demonstrate that SDMI successfully addresses the convergence challenges inherent in Meta-IS. Moreover, RVM-SDMI exhibits remarkable accuracy and computational efficiency in the system reliability analysis with small failure probabilities.

Meta-model based sequential importance sampling method for structural reliability analysis under high dimensional small failure probability

Reliability analysis poses a significant challenge for complex structures with stringent reliability requirements. While Sequential Importance Sampling (SIS) and Subset Simulation (SUS) have proven highly effective in addressing high-dimensional problems with small failure probabilities, the computational burden of mechanical simulations remains substantial due to the time-consuming nature of numerical simulation processes. Consequently, this paper introduces a novel approach, denoted as AK-SIS, which combines SIS with Kriging metamodeling specifically designed to address computational challenges associated with small failure probabilities. The fundamental principle of this approach involves utilizing AK-MCS technology (Echard et al., 2011) [1] as a precursor to the SIS approach to initially generate metamodels. These metamodels are then employed in lieu of performance functions in subsequent steps, significantly reducing the number of function calls required to simulate complex engineering problems when applying SIS techniques directly. By inheriting the advantages of SIS, AK-SIS has demonstrated its suitability for reliability analysis in scenarios involving high-dimensional spaces and small fault probabilities. Furthermore, AK-SIS is not limited by the shape of the failure domain, eliminates the need to solve the design point, and is particularly well-suited for analyzing reliability in cases of discontinuous failure domains, multiple failure domains, as well as complex failure domains and rare events. The efficacy of AK-SIS is substantiated through rigorous evaluation encompassing nonlinear, high-dimensional examples, and an engineering application. These empirical validations collectively contribute to a robust methodological framework for reliability analysis of intricate structures characterized by stringent reliability requirements.

Information Theoretical Importance Sampling Clustering and Its Relationship With Fuzzy C-Means

Information Theoretical Importance Sampling Clustering and Its Relationship With Fuzzy C-Means

A block-randomized stochastic method with importance sampling for CP tensor decomposition

A block-randomized stochastic method with importance sampling for CP tensor decomposition

Contrastive Continual Learning with Importance Sampling and Prototype-Instance Relation Distillation

Recently, because of the high-quality representations of contrastive learning methods, rehearsal-based contrastive continual learning has been proposed to explore how to continually learn transferable representation embeddings to avoid the catastrophic forgetting issue in traditional continual settings. Based on this framework, we propose Contrastive Continual Learning via Importance Sampling (CCLIS) to preserve knowledge by recovering previous data distributions with a new strategy for Replay Buffer Selection (RBS), which minimize estimated variance to save hard negative samples for representation learning with high quality. Furthermore, we present the Prototype-instance Relation Distillation (PRD) loss, a technique designed to maintain the relationship between prototypes and sample representations using a self-distillation process. Experiments on standard continual learning benchmarks reveal that our method notably outperforms existing baselines in terms of knowledge preservation and thereby effectively counteracts catastrophic forgetting in online contexts. The code is available at https://github.com/lijy373/CCLIS.

Fairness under Covariate Shift: Improving Fairness-Accuracy Tradeoff with Few Unlabeled Test Samples

Covariate shift in the test data is a common practical phenomena that can significantly downgrade both the accuracy and the fairness performance of the model. Ensuring fairness across different sensitive groups under covariate shift is of paramount importance due to societal implications like criminal justice. We operate in the unsupervised regime where only a small set of unlabeled test samples along with a labeled training set is available. Towards improving fairness under this highly challenging yet realistic scenario, we make three contributions. First is a novel composite weighted entropy based objective for prediction accuracy which is optimized along with a representation matching loss for fairness. We experimentally verify that optimizing with our loss formulation outperforms a number of state-of-the-art baselines in the pareto sense with respect to the fairness-accuracy tradeoff on several standard datasets. Our second contribution is a new setting we term Asymmetric Covariate Shift that, to the best of our knowledge, has not been studied before. Asymmetric covariate shift occurs when distribution of covariates of one group shifts significantly compared to the other groups and this happens when a dominant group is over-represented. While this setting is extremely challenging for current baselines, We show that our proposed method significantly outperforms them. Our third contribution is theoretical, where we show that our weighted entropy term along with prediction loss on the training set approximates test loss under covariate shift. Empirically and through formal sample complexity bounds, we show that this approximation to the unseen test loss does not depend on importance sampling variance which affects many other baselines.

DISCount: Counting in Large Image Collections with Detector-Based Importance Sampling

Many applications use computer vision to detect and count objects in massive image collections. However, automated methods may fail to deliver accurate counts, especially when the task is very difficult or requires a fast response time. For example, during disaster response, aid organizations aim to quickly count damaged buildings in satellite images to plan relief missions, but pre-trained building and damage detectors often perform poorly due to domain shifts. In such cases, there is a need for human-in-the-loop approaches to accurately count with minimal human effort. We propose DISCount -- a detector-based importance sampling framework for counting in large image collections. DISCount uses an imperfect detector and human screening to estimate low-variance unbiased counts. We propose techniques for counting over multiple spatial or temporal regions using a small amount of screening and estimate confidence intervals. This enables end-users to stop screening when estimates are sufficiently accurate, which is often the goal in real-world applications. We demonstrate our method with two applications: counting birds in radar imagery to understand responses to climate change, and counting damaged buildings in satellite imagery for damage assessment in regions struck by a natural disaster. On the technical side we develop variance reduction techniques based on control variates and prove the (conditional) unbiasedness of the estimators. DISCount leads to a 9-12x reduction in the labeling costs to obtain the same error rates compared to naive screening for tasks we consider, and surpasses alternative covariate-based screening approaches.

High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field

One crucial aspect of 3D head avatar reconstruction lies in the details of facial expressions. Although recent NeRF-based photo-realistic 3D head avatar methods achieve high-quality avatar rendering, they still encounter challenges retaining intricate facial expression details because they overlook the potential of specific expression variations at different spatial positions when conditioning the radiance field. Motivated by this observation, we introduce a novel Spatially-Varying Expression (SVE) conditioning. The SVE can be obtained by a simple MLP-based generation network, encompassing both spatial positional features and global expression information. Benefiting from rich and diverse information of the SVE at different positions, the proposed SVE-conditioned NeRF can deal with intricate facial expressions and achieve realistic rendering and geometry details of high-fidelity 3D head avatars. Additionally, to further elevate the geometric and rendering quality, we introduce a new coarse-to-fine training strategy, including a geometry initialization strategy at the coarse stage and an adaptive importance sampling strategy at the fine stage. Extensive experiments indicate that our method outperforms other state-of-the-art (SOTA) methods in rendering and geometry quality on mobile phone-collected and public datasets. Code and data can be found at https://github.com/minghanqin/AvatarSVE.

Machine-Learning-Based Prediction Modeling for Debris Flow Occurrence: A Meta-Analysis

Machine learning (ML) has become increasingly popular in the prediction of debris flow occurrence, but the various ML models utilized as baseline predictors reported in previous studies are typically limited to individual case bases. A comprehensive and systematic evaluation of existing empirical evidence on the utilization of ML as baseline predictors for debris flow occurrence is lacking. To address this gap, we conducted a meta-analysis of ML-based prediction modeling of debris flow occurrence by retrieving papers that were published between 2000 and 2023 from the Scopus and Web of Science databases. The general findings were as follows: (1) A total of 84 papers, distributed across 37 different journals in this time period, reflecting an overall upward trend. (2) Debris flow disasters occur throughout the world, and a total of 13 countries carried out research on the prediction of debris flow occurrence based on ML; China made significant contributions, but more research efforts in African countries should be considered. (3) A total of 36 categories of ML models were utilized as baseline predictors for debris flow occurrence, with logistic regression (LR) and random forest (RF) emerging as the most popular choices. (4) Feature engineering and model comparison were the most commonly utilized strategies in predicting debris flow occurrence based on ML (53 and 46 papers, respectively). (5) Interpretation methods were rarely utilized in predicting debris flow occurrence based on ML, with only 16 papers reporting their utilization. (6) In the prediction of debris flow occurrence based on ML, interpretation methods were rarely utilized, searching by data materials was the most important sample data source, the topographic factors were the most commonly utilized category of candidate variables, and the area under the ROC curve (AUROC) was the most frequently reported evaluation metric. (7) LR’s prediction performance for debris flow occurrence was inferior to that of RF, BPNN, and SVM; SVM was comparable to RF, and all superior to BPNN. (8) The application process for the prediction of debris flow occurrence based on ML consisted of three main steps: data preparation, model construction and evaluation, and prediction outcomes. The research gaps in predicting debris flow occurrence based on ML include utilizing new ML techniques and enhancing the interpretability of ML. Consequently, this study contributes both to academic ML research and to practical applications in the prediction of debris flow occurrence.

Abstract 6101: PLASMUT: An R Package for estimating the probability of tumor-specific mutations in cell-free DNA

Abstract Noninvasive approaches for detection of tumor-specific mutations in cell-free DNA (cfDNA) have the potential to track a patient’s response to treatment, enabling effective and timely decisions on therapy. However, mutations in cfDNA arising from clonal hematopoeisis (CH) are common and tumor biopsies for definitive identification of the origin of these mutations are not always available. Sequencing of matched cells from buffy coat and the absence of mutations in these cells has been used to rule-out white blood cell (WBC) mutations, but uneven sequencing depths between matched cfDNA and buffy coat and the fraction of mutant alleles are generally ignored by rule-based tests. A probabilistic approach that quantifies the evidence of tumor-derived mutations in cfDNA is needed. We developed a Bayesian framework to estimate the probability that a mutation identified in cfDNA is tumor specific. Our approach requires the number of reads with a mutant allele in plasma (yp) and WBCs (yw) and the corresponding number of total distinct reads at these locations. The posterior odds that a mutation is tumor-derived (S) versus CH or germline (H) is given by the ratio of the probabilities of observing the distinct reads given each model times the prior odds. Estimation of the Bayes factor is obtained by integrating over the unobserved mutant allele fractions in plasma and WBCs using Monte Carlo importance sampling. We applied this approach to 52 patients with initially unresectable colorectal cancer (CRC) liver metastases in the CAIRO5 clinical trial (NCT02162563), using ultra-deep targeted sequencing of cfDNA from plasma and matched WBCs. Among the CAIRO5 patients analyzed, we identified 95 mutations with moderate evidence of tumor-derived cfDNA mutations (Bayes factor > 10) and 19 mutations that were CH-derived (Bayes factor < 0.1). For a subset of 47 cfDNA mutations with no corresponding mutation identified by WBC sequencing, the evidence of tumor origin was highly variable (Bayes factor range: 0.03 to 5.6). While the standard rule-based approach identifies all of these mutations as tumor-derived, none of these mutations reach a moderate evidence cutpoint (all Bayes factors < 6). As a false positive would lead to identification of cfDNA mutations that do not track tumor burden, requiring even higher levels of evidence (Bayes factor > 99) for the selection of cfDNA mutations could be warranted, and still identifies one or more cfDNA mutations in 43 of the patients. This approach is implemented in the R package PLASMUT available from Bioconductor (doi:10.18129/B9.bioc.plasmut). We developed an approach that quantifies the evidence between two competing models for the origin of mutations in cfDNA. A cutpoint for determination of the probability of tumor-derived cfDNA mutations can be tailored to the disease application, balancing the potential benefits of noninvasive testing with the harms of false positives and negatives. Citation Format: Adith S. Arun, Jamie E. Medina, Stephen Cristiano, Daniel C. Bruhm, Remond J. Fijneman, Gerrit A. Meijer, Alessandro Leal, Victor E. Velculescu, Robert B. Scharpf. PLASMUT: An R Package for estimating the probability of tumor-specific mutations in cell-free DNA [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6101.

Meta model-based and cross entropy-based importance sampling algorithms for efficiently solving system failure probability function

The multi-mode system failure probability function (SFPF) can quantify how the distribution parameters of the random input vector affect the system safety and decouple the system reliability-based design optimization model. However, for a problem with a time-consuming implicit performance function and rare failure domain, efficiently solving the SFPF remains significantly challenging. Therefore, in this study, two efficient algorithms are proposed, namely, the meta model-based importance sampling and cross entropy-based importance sampling. The contributions of this study are twofold. The first is constructing a single-loop optimal importance sampling density (SL-OISD) method to decouple the double-loop framework for analyzing the SFPF. The second is establishing two methods to efficiently approximate the SL-OISD and complete the SFPF estimation. The first method is based on the meta model of the system performance function, which is abbreviated as SL-Meta-IS. The second method is based on minimizing the cross entropy between the Gaussian mixture density model and SL-OISD, which is abbreviated as SL-CE-IS. To reduce the number of evaluating the system performance function when approximating the SL-OISD, sampling the SL-OISD, and identifying the state of the samples for completing the SFPF estimation, an adaptive Kriging model of the system performance function is introduced into SL-Meta-IS and SL-CE-IS. Owing to decoupling the double-loop framework into a single-loop framework, replacing the time-consuming system performance function with the economic Kriging model, and employing importance sampling variance reduction techniques to address issues related to the rare failure domain, the proposed SL-Meta-IS and SL-CE-IS methods greatly enhance the efficiency of SFPF estimations. The numerical and practical examples demonstrate that the two proposed methods are superior to the existing algorithms; moreover, the efficiency of SL-CE-IS is higher than that of SL-Meta-IS.

Deep reinforcement learning based on balanced stratified prioritized experience replay for customer credit scoring in peer-to-peer lending

In recent years, deep reinforcement learning (DRL) models have been successfully utilised to solve various classification problems. However, these models have never been applied to customer credit scoring in peer-to-peer (P2P) lending. Moreover, the imbalanced class distribution in experience replay, which may affect the performance of DRL models, has rarely been considered. Therefore, this article proposes a novel DRL model, namely a deep Q-network based on a balanced stratified prioritized experience replay (DQN-BSPER) model, for customer credit scoring in P2P lending. Firstly, customer credit scoring is formulated as a discrete-time finite-Markov decision process. Subsequently, a balanced stratified prioritized experience replay technology is presented to optimize the loss function of the deep Q-network model. This technology can not only balance the numbers of minority and majority experience samples in the mini-batch by using stratified sampling technology but also select more important experience samples for replay based on the priority principle. To verify the model performance, four evaluation measures are introduced for the empirical analysis of two real-world customer credit scoring datasets in P2P lending. The experimental results show that the DQN-BSPER model can outperform four benchmark DRL models and seven traditional benchmark classification models. In addition, the DQN-BSPER model with a discount factor γ of 0.1 has excellent credit scoring performance.

Simulating Z2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$Z_{2}$\\end{document} lattice gauge theory with the variational quantum thermalizer

The properties of strongly-coupled lattice gauge theories at finite density as well as in real time have largely eluded first-principles studies on the lattice. This is due to the failure of importance sampling for systems with a complex action. An alternative to evade the sign problem is quantum simulation. Although still in its infancy, a lot of progress has been made in devising algorithms to address these problems. In particular, recent efforts have addressed the question of how to produce thermal Gibbs states on a quantum computer. In this study, we apply a variational quantum algorithm to a low-dimensional model which has a local abelian gauge symmetry. We demonstrate how this approach can be applied to obtain information regarding the phase diagram as well as unequal-time correlation functions at non-zero temperature.

Bayesian forecasting of stock returns on the JSE using simultaneous graphical dynamic linear models*

ABSTRACT Cross-series dependencies are crucial in obtaining accurate forecasts when forecasting a multivariate time series. Simultaneous Graphical Dynamic Linear Models (SGDLMs) are Bayesian models that elegantly capture cross-series dependencies. This study forecasts returns of a 40 -dimensional time series of stock data from the Johannesburg Stock Exchange (JSE) using SGDLMs. The SGDLM approach involves constructing a customised dynamic linear model (DLM) for each univariate time series. At each time point, the DLMs are recoupled using importance sampling and decoupled using mean-field variational Bayes. Our findings indicate that the SGDLM forecasts stock data similarly to decoupled DLMs, as evident from values of root mean square errors and mean absolute deviations. The SGDLM forecasts the returns quite accurately as seen from the slightly overestimated prediction interval coverages.