Computational Statistics

Bayesian Computational Methods for High-Dimensional Data

Development of scalable Bayesian inference algorithms for problems involving thousands or millions of parameters using variational approximation and sampling techniques.

RESEARCH GAP FRONTIERS

Gradient-Free Optimization in Complex Landscapes

Design and analysis of optimization algorithms that operate without gradient information for non-differentiable, multimodal, or black-box objective functions.

RESEARCH GAP FRONTIERS

Probabilistic Graphical Models Inference

Computational methods for exact and approximate inference in Bayesian networks, Markov random fields, and factor graphs.

RESEARCH GAP FRONTIERS

Approximate Bayesian Computation for Intractable Likelihoods

Statistical inference techniques for simulator-based models where likelihood functions are computationally intractable or unavailable.

RESEARCH GAP FRONTIERS

Markov Chain Monte Carlo Convergence Analysis

Theoretical and empirical study of mixing times, convergence diagnostics, and efficiency measures for MCMC sampling algorithms.

RESEARCH GAP FRONTIERS

Variational Inference with Neural Networks

Integration of deep learning architectures with variational inference frameworks for scalable approximate posterior computation.

RESEARCH GAP FRONTIERS

Importance Sampling Methods and Adaptive Schemes

Development of adaptive importance sampling distributions and reweighting techniques to reduce variance in Monte Carlo estimation.

RESEARCH GAP FRONTIERS

Distributed Statistical Computing Algorithms

Parallel and distributed implementations of statistical algorithms across multiple processors or computing nodes for massive datasets.

Stochastic Gradient Descent Convergence Theory

Mathematical analysis of convergence rates, generalization bounds, and optimization behavior of stochastic gradient methods.

Sequential Monte Carlo Methods and Particle Filters

Computational techniques for state-space models and online inference using sequential importance sampling with resampling.

Hamiltonian Monte Carlo and Geometric Methods

Advanced MCMC samplers utilizing differential geometry and Hamiltonian dynamics for efficient exploration of complex posterior distributions.

Empirical Risk Minimization and Regularization

Computational methods for solving penalized optimization problems with analysis of statistical learning theory and generalization.

Kernel Methods and Gaussian Process Computation

Scalable algorithms for kernel machines and Gaussian process inference including sparse approximations and structure exploitation.

Expectation-Maximization Algorithm Extensions

Development of variants and accelerations of EM for missing data problems, mixture models, and latent variable models.

Computational Challenges in Causal Inference

Algorithms and software for estimating causal effects under complex experimental designs, unmeasured confounding, and interference.

Bootstrap Methods for Uncertainty Quantification

Computational resampling techniques for constructing confidence intervals, hypothesis tests, and approximating sampling distributions.

Sparse Linear Regression and Feature Selection

Computational algorithms for high-dimensional linear models using regularization techniques like lasso, ridge, and elastic net.

Matrix Factorization and Tensor Decomposition

Algorithms for low-rank approximations, dimensionality reduction, and structure discovery in matrix and multi-dimensional data.

Reinforcement Learning with Statistical Guarantees

Computational methods for sequential decision-making with theoretical analysis of sample complexity and regret bounds.

Online Learning and Streaming Data Algorithms

Algorithms for statistical learning from data streams with memory constraints and single-pass processing requirements.

Variational Autoencoder Optimization Techniques

Computational methods for training variational autoencoders including reparameterization tricks and gradient estimation.

Generative Adversarial Network Training Stability

Analysis and improvement of computational stability, convergence properties, and mode coverage in adversarial training.

Sampling from Complex Energy-Based Models

Computational techniques for sampling from unnormalized distributions including Boltzmann machines and neural energy models.

Latent Dirichlet Allocation and Topic Modeling

Efficient inference algorithms for discovering latent topics in document collections using variational and sampling methods.

Nonparametric Bayesian Computation and Inference

Computational algorithms for Dirichlet processes, Gaussian processes, and other nonparametric priors with unknown model complexity.

Optimal Transport and Wasserstein Distances

Computational methods for computing optimal transport plans and using Wasserstein metrics in statistical inference.

Neural Network Training Dynamics and Theory

Mathematical analysis of neural network optimization landscapes, implicit regularization, and generalization phenomena.

Quantile Regression Computation and Algorithms

Efficient computational methods for quantile regression including algorithms for large-scale problems and nonparametric variants.

Survival Analysis Computational Methods

Algorithms for censored data analysis, Cox models, competing risks, and computational survival model fitting.

Spatial Statistics and Kriging Computation

Scalable algorithms for spatial prediction, kriging interpolation, and estimation in geostatistical applications.

Time Series Forecasting with Machine Learning

Computational methods for temporal prediction including ARIMA, state-space models, and neural network approaches.

High-Dimensional Multiple Testing Correction

Computational procedures for controlling false discovery rates and family-wise error rates in large-scale hypothesis testing.

Mixture Model Estimation and Selection

Algorithms for fitting finite and infinite mixture models, determining optimal number of components, and model comparison.

Cross-Validation and Model Selection Methods

Computational strategies for hyperparameter tuning, model selection, and estimation of prediction error in statistical learning.

Permutation Testing and Resampling Inference

Computational nonparametric hypothesis testing methods that do not require distributional assumptions.

Robust Statistics and Outlier Detection

Algorithms for robust parameter estimation resistant to outliers and contamination in statistical data analysis.

Bayesian Model Averaging Computation

Computational methods for combining predictions and inference across multiple models weighted by posterior model probabilities.

Semi-Supervised Learning Algorithms

Computational methods leveraging labeled and unlabeled data simultaneously for improved statistical learning with limited supervision.

Transfer Learning and Domain Adaptation

Computational techniques for adapting statistical models across different distributions, domains, or tasks.

Federated Learning and Privacy-Preserving Statistics

Algorithms for distributed learning across decentralized data sources while maintaining privacy and communication efficiency.

Information Geometry and Statistical Manifolds

Application of differential geometric methods to understand statistical models, divergences, and optimization on parameter manifolds.

Natural Gradient Methods in Statistics

Optimization algorithms utilizing Fisher information and information geometric structure for improved convergence.

Influence Functions and Sensitivity Analysis

Computational methods for assessing how data points influence statistical estimates and detecting influential observations.

Clustering Algorithms and Computational Complexity

Analysis and development of scalable clustering methods including k-means, hierarchical clustering, and density-based approaches.

Regression Trees and Ensemble Methods

Algorithms for decision tree construction, pruning, bagging, random forests, and gradient boosting for regression.

Classification Trees and Recursive Partitioning

Computational methods for binary and multi-class classification using tree-based algorithms and ensemble averaging.

Anomaly Detection in High Dimensions

Algorithms for identifying outliers, unusual patterns, and anomalies in high-dimensional data for applications in monitoring.

Graphical Lasso and Sparse Precision Matrices

Computational methods for estimating sparse inverse covariance matrices and discovering conditional independence structures.

Principal Component Analysis and Extensions

Algorithms for dimensionality reduction via orthogonal projection including sparse PCA, robust PCA, and kernelized variants.

Differentiable Programming for Statistical Inference

Development and analysis of automatic differentiation techniques for computing gradients in complex statistical models and Bayesian inference algorithms.

Scalable Variational Inference for Massive Datasets

Computational methods for performing variational inference on datasets too large for traditional approaches through stochastic optimization and data subsampling.

Normalizing Flows for Density Estimation

Theory and computation of invertible neural network transformations for learning complex probability distributions and performing efficient sampling.

Score-Based Generative Models and Diffusion Processes

Development of computational techniques for training generative models via score function estimation and reverse-time stochastic differential equations.

Contrastive Divergence and Energy-Based Learning

Algorithms and convergence analysis for training energy-based models through contrastive learning objectives and gradient-based sampling approximations.

Invertible Neural Networks for Bijective Mappings

Design and computation of neural network architectures maintaining invertibility for generative modeling, density estimation, and variational inference.

Bayesian Deep Learning with Uncertainty Quantification

Computational methods for obtaining posterior distributions over neural network parameters and propagating uncertainty through deep learning models.

Gibbs Sampling for Discrete Latent Variable Models

Efficient sampling algorithms and convergence analysis for probabilistic models with discrete latent variables and categorical distributions.

Langevin Dynamics and Gradient-Based Sampling

Computational theory and algorithms for sampling from posterior distributions using overdamped and underdamped Langevin dynamics with convergence guarantees.

Variational Graph Autoencoders and Network Inference

Development of computational methods for learning latent representations of graphs and networks through variational inference frameworks.

Expectation Propagation and Message Passing Algorithms

Advanced message passing algorithms for approximate inference in probabilistic graphical models with analysis of convergence and accuracy.

Approximate Inference via Belief Propagation

Computational methods and convergence theory for sum-product and max-product algorithms in graphical models with applications to inference problems.

Neural Ordinary Differential Equations and Learning Dynamics

Computational techniques for training neural networks parameterized as continuous dynamical systems with applications to density estimation and generative modeling.

Sliced Wasserstein Distance Approximation

Efficient computational methods for approximating optimal transport distances through random projections for scalable statistical inference.

Neural Tangent Kernel Theory and Overparameterization

Theoretical and computational analysis of infinite-width neural network limits and their connection to kernel methods in statistical learning.

Copula-Based Statistical Dependency Modeling

Computational algorithms for estimating and sampling from copula models capturing complex multivariate dependencies with applications to risk assessment.

Meta-Learning and Few-Shot Statistical Inference

Computational frameworks for learning statistical models that adapt rapidly to new data distributions with minimal training samples.

Causal Forest Methods for Treatment Effect Estimation

Development and analysis of tree-based algorithms for estimating heterogeneous treatment effects and causal relationships from observational data.

Debiased Machine Learning for Policy Evaluation

Computational methods for combining flexible machine learning models with debiasing techniques to obtain valid statistical inference for policy effects.

Autoencoder-Based Representation Learning Theory

Analysis of what representations autoencoders learn, when they recover meaningful statistical structure, and guarantees for downstream inference tasks.

Extreme Value Statistics and Tail Risk Computation

Computational methods for modeling and inference in extreme value distributions with applications to rare events and tail risk assessment.

Hidden Markov Models and State Space Inference

Efficient filtering, smoothing, and parameter estimation algorithms for sequential latent variable models with theoretical convergence analysis.

Optimal Experimental Design and Active Learning

Computational algorithms for selecting informative data points and designing experiments to maximize statistical efficiency and information gain.

Gradient Boosting Machines and Additive Models

Theoretical analysis and computational optimization of boosting algorithms for fitting flexible additive models through greedy optimization.

Sparsity-Inducing Priors and Computational Methods

Bayesian computational techniques for inference with spike-and-slab and horseshoe priors promoting sparse solutions in high dimensions.

Posterior Sampling via Data Augmentation Schemes

MCMC algorithms exploiting latent variable augmentation strategies to improve mixing and computational efficiency for intractable posteriors.

Statistical Guarantees for Deep Reinforcement Learning

Theoretical analysis providing sample complexity bounds and convergence guarantees for value and policy learning in high-dimensional state spaces.

Surrogate Models and Emulation for Expensive Simulations

Computational methods for building fast statistical approximations to expensive computer simulations for uncertainty quantification and optimization.

Reverse Engineering via Statistical Causal Discovery

Algorithms and statistical theory for learning directed acyclic graph structures from observational data under causal assumptions.

Approximate Message Passing and State Evolution

Theoretical analysis of iterative algorithms in high dimensions through state evolution equations predicting performance of approximate message passing schemes.

Functional Data Analysis and Curve Registration

Computational methods for statistical inference on infinite-dimensional functional data including smoothing, alignment, and functional linear regression.

Multi-Task Learning with Parameter Sharing

Computational frameworks for learning shared representations and parameters across related statistical tasks improving overall prediction and generalization.

Doubly Robust Estimation and Orthogonal Learning

Development of computational methods providing protection against nuisance parameter estimation errors in semiparametric inference problems.

Poisson Approximation and Chen-Stein Methods

Theoretical and computational techniques for bounding approximation errors in rare event probabilities and count data inference.

Attention Mechanisms for Statistical Modeling

Theory and computation of attention-based architectures for learning structured dependencies and interpretable statistical relationships in data.

Implicit Likelihood Models and Score Matching

Computational inference methods for models where likelihood evaluation is intractable using score matching and related divergence minimization approaches.

Randomized Linear Algebra for Statistical Computation

Probabilistic algorithms for matrix operations achieving speedups through random sketching with rigorous statistical error analysis.

Interacting Particle Systems and Mean-Field Approximations

Theoretical analysis of many-particle systems in statistical mechanics with applications to Monte Carlo methods and mean-field approximations.

Subsampling for Scalable Bayesian Inference

Computational techniques using data subsampling and likelihood scaling for Bayesian posterior inference on massive datasets.

Graphical Model Structure Learning and Model Selection

Algorithms for learning sparse network structures underlying data distributions with theoretical guarantees for correct edge recovery.

Stochastic Variational Inference with Natural Gradient

Computational optimization of variational objectives using natural gradient descent for efficient large-scale Bayesian inference.

Conformal Prediction and Distribution-Free Inference

Methods for constructing prediction sets and confidence regions with finite-sample coverage guarantees without distributional assumptions.

Mixture of Experts and Computational Scaling

Algorithms for training conditional mixture models with expert-gating networks for efficient learning from heterogeneous data.

ABC Shadow Filtering for Partially Observable Systems

Computational methods for Bayesian inference in partially observed models combining approximate Bayesian computation with particle filtering.

Annealed Importance Sampling and Tempering Methods

Sequential sampling algorithms using temperature schedules to bridge between easy and target distributions for efficient exploration.

Smoothed Particle Inference and Likelihood-Free Methods

Computational approaches for Bayesian inference when likelihoods are unavailable using particle-based approximations and smoothing kernels.

Compositional Data Analysis and Log-Ratio Transformations

Statistical methods and computation for data constrained to simplices including proper centering and transformations for valid inference.

Cyclical Learning Rates and Optimizer Tuning

Theoretical and empirical analysis of adaptive learning rate schedules for stochastic optimization with applications to neural network training.

Geometric Ergodicity and MCMC Diagnostics

Rigorous analysis of Markov chain mixing properties and development of computational diagnostic tools for assessing convergence to stationarity.

Information Bottleneck and Compression Bounds

Theory of learning under information constraints with computational implications for neural network training and generalization bounds.

Scalable Expectation Propagation for Massive Datasets

Development of efficient expectation propagation algorithms that maintain accuracy while processing datasets with millions of observations and thousands of features.

Adaptive Metropolis-Hastings with Machine Learning Proposals

Integration of neural networks and other machine learning models to generate adaptive proposal distributions in Metropolis-Hastings algorithms.

Computational Methods for Copula-Based Inference

Development of efficient algorithms for estimating and performing inference with high-dimensional copula models for multivariate dependence.

Sparse Precision Matrix Estimation and Graph Learning

Computational techniques for learning sparse graphical models through efficient precision matrix estimation under various sparsity assumptions.

Variational Bounds for Intractable Likelihood Models

Development of tight variational lower bounds and upper bounds for models with intractable likelihood functions using novel optimization techniques.

Parallel Tempering and Replica Exchange Methods

Advanced computational strategies for parallel tempering algorithms that improve mixing and convergence in multimodal posterior distributions.

Stochastic Optimization for Non-Convex Statistical Problems

Analysis and development of stochastic algorithms for non-convex optimization arising in statistical estimation and machine learning.

Approximate Inference in Continuous Markov Random Fields

Computational methods for approximating inference in continuous-valued Markov random fields and related continuous graphical models.

Spectral Methods for Statistical Learning and Estimation

Application of spectral decomposition and eigenvalue-based techniques for scalable parameter estimation and unsupervised learning.

Variance Reduction Techniques in Monte Carlo Methods

Development of advanced variance reduction techniques including control variates and antithetic sampling for improved Monte Carlo efficiency.

Computational Topology and Persistent Homology in Statistics

Application of topological data analysis and persistent homology algorithms for statistical inference and shape analysis.

Multivariate Functional Data Analysis Algorithms

Computational methods for analyzing and inferring from high-dimensional functional data including curves and surfaces.

Accelerated Proximal Methods for Regularized Estimation

Development of accelerated proximal gradient and splitting methods for efficiently solving large-scale regularized statistical problems.

Computational Inference for Point Process Models

Efficient algorithms for likelihood evaluation, parameter estimation, and prediction in spatial and temporal point process models.

Sequential Anomaly Detection and Change Point Analysis

Online computational methods for detecting anomalies and structural changes in streaming data with statistical guarantees.

Laplace Approximation and Higher-Order Methods

Development of improved Laplace approximations and higher-order Taylor expansions for approximate Bayesian inference.

Neural Operator Learning for Statistical Emulation

Learning neural operators that can approximate complex statistical models and computationally expensive likelihood functions.

Computational Methods for Latent Variable Models

Development of efficient inference algorithms for models with discrete and continuous latent variables including mixture models.

GPU-Accelerated Statistical Algorithms and Libraries

Implementation and optimization of statistical algorithms on graphics processing units for massive acceleration of computations.

Normalizing Flows for Flexible Posterior Approximation

Development of flexible normalizing flow models for accurate approximation of complex posterior distributions in Bayesian inference.

Composite Likelihood and Pairwise Likelihood Methods

Computational techniques for inference using composite likelihoods when full likelihood evaluation is intractable or expensive.

Manifold Learning and Dimensionality Reduction Algorithms

Development of computational methods for discovering low-dimensional manifolds and performing inference on manifold-constrained data.

Convex Relaxations for Combinatorial Statistical Problems

Design of convex relaxations and semidefinite programming approaches for intractable combinatorial inference problems.

Bayesian Optimization for Expensive Computer Experiments

Efficient algorithms for Bayesian optimization when objective function evaluations are costly using Gaussian processes and acquisition functions.

Computational Methods for Networked Data Analysis

Statistical algorithms for inference on network and graph-structured data including community detection and network regression.

Score Matching and Contrastive Learning Methods

Computational techniques using score matching and contrastive objectives for training energy-based models without partition functions.

Tensor Network and Belief Propagation Algorithms

Utilization of tensor network methods and belief propagation for efficient inference in high-dimensional graphical models.

Data Augmentation Strategies for Statistical Computation

Development of effective data augmentation techniques to improve computational efficiency and convergence in statistical algorithms.

Gradient-Based Sampling with Score Functions

Computational methods that exploit gradient information from score functions for improved sampling efficiency in Bayesian inference.

Coupling Methods for Probability Distance Estimation

Development of efficient coupling and transportation-based methods for computing distances between probability distributions.

Multi-Task Learning and Meta-Learning Algorithms

Computational approaches for learning shared representations across multiple related statistical tasks and domains.

Bayesian Nonparametrics with Scalable Inference

Development of scalable inference algorithms for Bayesian nonparametric models including Dirichlet processes and Indian buffet processes.

Computational Statistics for Differential Equations

Statistical inference methods for models defined by ordinary and partial differential equations using computational techniques.

Slice Sampling and Adaptive Slice Schemes

Development of efficient slice sampling algorithms with adaptive schemes that improve mixing and reduce tuning requirements.

Causal Discovery through Constraint-Based Methods

Computational algorithms for discovering causal structure from observational data using constraint-based approaches.

Minibatch Statistics and Subsampling Methods

Theoretical analysis and development of statistical methods that maintain validity when computed on subsamples of data.

Gaussian Approximations and Mean Field Methods

Development of improved Gaussian approximations and mean-field variational methods for complex posterior inference.

Computational Methods for Survival and Event History Data

Efficient algorithms for maximum likelihood and Bayesian estimation in survival analysis and event history models.

Neural Density Ratio Estimation and Classification

Learning density ratios between distributions using neural networks for likelihood-free inference and importance weighting.

Computational Methods for Mixed Effects Models

Development of scalable algorithms for estimation and inference in linear and nonlinear mixed effects models.

Risk Minimization and Empirical Process Theory

Theoretical foundations and algorithms for estimating risk and learning bounds through empirical process analysis.

Computational Statistics for Image and Signal Analysis

Statistical algorithms for processing and inference in image and signal data including denoising and reconstruction.

Quasi-Monte Carlo Methods and Randomization

Development of quasi-Monte Carlo techniques with low-discrepancy sequences for improved convergence in numerical integration.

Diffusion-Based Generative Models and Sampling

Computational methods for training and sampling from diffusion-based generative models for statistical inference.

Computational Statistics for Genomic Data Analysis

Development of efficient algorithms for statistical analysis of high-dimensional genomic and biological data.

Amortized Inference with Conditional Neural Networks

Techniques for learning amortized inference networks that quickly perform Bayesian inference for multiple observations.

Coordinate Descent and Block Optimization Methods

Development of efficient coordinate descent and block optimization algorithms for high-dimensional statistical problems.

Computational Methods for Count and Categorical Data

Specialized algorithms for efficient inference and estimation in models for discrete and categorical response variables.

Probabilistic Program Inference and Synthesis

Computational methods for Bayesian inference in probabilistic programming languages and program synthesis applications.

Adaptive Metropolis-Hastings and Convergence Diagnostics

Investigation of self-tuning MCMC algorithms with real-time convergence assessment and automatic stopping criteria for complex posterior distributions.

Computational Methods for Causal Graphs Discovery

Algorithm development for learning directed acyclic graphs and causal structures from observational data with computational efficiency constraints.

Differentiable Monte Carlo Estimators

Design of Monte Carlo sampling schemes whose estimates are differentiable with respect to model parameters for gradient-based optimization.

Approximate Inference in Continuous-Time Models

Computational approaches for parameter estimation and uncertainty quantification in stochastic differential equations and point processes.

Score-Based Generative Modeling Algorithms

Development of score matching and diffusion-based computational techniques for learning and sampling from high-dimensional data distributions.

Debiasing Stochastic Gradient Estimates

Methods for correcting bias in minibatch gradient estimates to achieve optimal convergence rates in online statistical learning.

Continuous Normalizing Flows and ODE Solvers

Integration of neural ordinary differential equations with probabilistic modeling for flexible density estimation and likelihood computation.

Stein Variational Descent Methods

Particle-based variational inference using Stein discrepancies to generate samples that approximate posterior distributions efficiently.

Accelerated Proximal Algorithms for Statistics

Development of fast first-order methods combining acceleration and proximal operators for non-smooth statistical optimization problems.

Computationally Efficient Model Checking Methods

Scalable posterior predictive checking and goodness-of-fit testing techniques that avoid expensive full model simulations.

Selective Inference and Conditioning Algorithms

Computational methods for valid inference after model selection, variable screening, and data-dependent procedures.

Randomized Matrix Algorithms for Linear Regression

Fast sketching and random projection techniques for solving large-scale least squares problems with theoretical accuracy guarantees.

Consensus and Distributed MCMC Methods

Decentralized Bayesian inference algorithms where multiple agents cooperatively sample from global posteriors without centralized communication.

Scalable Expectation Propagation Algorithms

Efficient message-passing algorithms for approximate Bayesian inference in factor graphs with applications to large-scale problems.

Gradient Flow Dynamics and Implicit Bias

Analysis of continuous-time gradient flows to understand implicit regularization effects and generalization in statistical learning.

Computational Topology and Persistent Homology

Application of topological data analysis techniques for feature extraction and inference in high-dimensional statistical contexts.

Variational Approximations for Mixed Effects Models

Scalable inference algorithms for hierarchical and mixed-effects models using variational methods tailored to random effect structures.

Reparameterization Tricks for Discrete Latent Variables

Gradient estimators and computational schemes for learning discrete latent variable models using continuous relaxations.

Spectral Methods for Nonparametric Statistics

Eigenvalue and singular value decomposition based approaches for efficient computation in nonparametric density and function estimation.

Computational Inference in Epidemic Models

Algorithms for parameter inference and prediction in stochastic compartmental disease models with incomplete observation data.

GPU-Accelerated Statistical Computing

Design and implementation of statistical algorithms exploiting GPU parallelism for massive speedups in Bayesian and frequentist methods.

Splitting and Alternating Direction Methods

Development of operator splitting algorithms for convex and non-convex statistical optimization with convergence guarantees.

Empirical Likelihood Computation and Optimization

Efficient algorithms for computing empirical likelihood ratios and conducting hypothesis tests in high-dimensional settings.

Neural Ordinary Differential Equations for Inference

Applications of neural ODE architectures to statistical inference problems requiring continuous-time dynamics modeling.

Minimax Optimal Algorithms and Lower Bounds

Development and proof of computational algorithms achieving statistical minimax rates with information-theoretic lower bound verification.

Bayesian Optimization and Active Learning

Algorithms for sequential experimentation and adaptive design of statistical experiments using surrogate models and acquisition functions.

Computational Methods for Functional Data Analysis

Algorithms for inference with infinite-dimensional function-valued data including smoothing, registration, and dimension reduction.

Scalable Gaussian Process Approximations

Inducing point, low-rank, and local methods for computational feasibility of Gaussian processes on large datasets.

Computation Under Privacy Constraints

Statistical algorithms with differential privacy guarantees achieving inference objectives while bounding disclosure risk.

Path Sampling and Thermodynamic Integration

Computational methods for model evidence and marginal likelihood estimation using integration along interpolating paths.

Quantile-Based and Robust Optimization Methods

Algorithms for statistical inference using robust loss functions and quantile-based objectives resistant to outliers.

Iteratively Reweighted Algorithms for Sparsity

Computational schemes for non-convex sparse estimation using iterative reweighting strategies with convergence guarantees.

Sampling from Intractable Densities via Coupling

Coupling-based MCMC methods for exact inference without tractability assumptions on intractable target densities.

Computational Approaches to Time-Varying Networks

Algorithms for dynamic network inference and evolution tracking with streaming data and evolving edge structures.

High-Dimensional Covariance Estimation and Shrinkage

Computational methods for stable covariance matrix estimation in ultra-high dimensions using shrinkage and regularization.

Doubly Robust Computation and Semiparametric Inference

Algorithms for semiparametric efficiency achieving parametric convergence rates under weak nuisance parameter estimates.

Simulation-Based Calibration and Diagnostics

Methods for validating Bayesian computational algorithms through simulation studies with coverage and efficiency assessment.

Data Augmentation and Latent Variable Algorithms

Computational techniques exploiting augmented problem structures to accelerate inference in high-dimensional latent variable models.

Convex Relaxations and Semidefinite Programming

Formulation of non-convex statistical problems as convex relaxations solvable via semidefinite programming with quality guarantees.

Online Convex Optimization for Statistics

Algorithms for sequential decision making and parameter estimation using online learning theory with regret bounds.

Computational Inference in Matching and Causality

Scalable algorithms for optimal matching, balance checking, and sensitivity analysis in observational causal studies.

Quasi-Monte Carlo Methods in Bayesian Inference

Low-discrepancy sequence-based integration for improved convergence rates in Bayesian posterior computation and approximation.

Structure-Exploiting Optimization for Graphical Models

Algorithms leveraging sparsity and decomposition structure of graphical models for efficient learning and inference.

Scalable Inference for Massive Graph Networks

Development of distributed computational algorithms for Bayesian and frequentist inference on large-scale network data with billions of nodes and edges.

Computational Methods for Categorical Data Analysis

Algorithms for inference and prediction with categorical and count response variables including log-linear models.

Adaptive Monte Carlo for Multimodal Posteriors

Design of computationally efficient sampling schemes that dynamically adapt to complex posterior landscapes with multiple isolated modes and varying geometry.

Second-Order Methods and Newton Algorithms

Development of Newton and quasi-Newton methods for statistical optimization with applications to high-dimensional problems.

Computational Aspects of Multiple Hypothesis Testing

Algorithms for controlling error rates in multiple testing with computational efficiency and power optimization.

Debiased Machine Learning and Causal Effect Estimation

Computational methods for extracting causal parameters from high-dimensional nuisance models while maintaining statistical efficiency and asymptotic normality.

Laplace Approximations and Saddle Point Methods

Computational schemes using Laplace approximation and saddle point expansions for likelihood and partition function inference.

Differentiable Probabilistic Programming Languages

Development of computational frameworks that combine automatic differentiation with probabilistic inference to enable efficient gradient-based Bayesian computation.

Computational Optimal Experimental Design

Algorithms for real-time optimization of experimental designs under computational constraints using information-theoretic criteria and sequential decision-making.