A Conceptual Introduction to Hamiltonian Monte Carlo
read the original abstract
Hamiltonian Monte Carlo has proven a remarkable empirical success, but only recently have we begun to develop a rigorous understanding of why it performs so well on difficult problems and how it is best applied in practice. Unfortunately, that understanding is confined within the mathematics of differential geometry which has limited its dissemination, especially to the applied communities for which it is particularly important. In this review I provide a comprehensive conceptual account of these theoretical foundations, focusing on developing a principled intuition behind the method and its optimal implementations rather of any exhaustive rigor. Whether a practitioner or a statistician, the dedicated reader will acquire a solid grasp of how Hamiltonian Monte Carlo works, when it succeeds, and, perhaps most importantly, when it fails.
This paper has not been read by Pith yet.
Forward citations
Cited by 60 Pith papers
-
Establishing an $\Omega(\sqrt{d})$ complexity lower bound for PDMP samplers and how to break it: a sub-$\sqrt{d}$ algorithm for Gaussian-tailed targets
Proves Ω(√d) lower bound for PDMP samplers and presents a new adaptive PDMP achieving sub-√d empirical complexity O(d^α) (α∈[0.2,0.3]) for Gaussian-tailed targets.
-
A Theoretical Comparison of No-U-Turn Sampler Variants: Necessary and Sufficient Convergence Conditions and Mixing Time Analysis under Gaussian Targets
NUTS-mul and NUTS-BPS show nearly identical qualitative ergodicity behavior depending on target tails, with both mixing in O(d^{1/4}) time for Gaussians but smaller constants for NUTS-BPS.
-
Calibration, Not Compilation: Detecting and Repairing Misspecified Probabilistic Programs Written by Language Models
Bayesian workflow diagnostics outperform unit tests for detecting and repairing statistically misspecified LLM-generated probabilistic programs across benchmarks and real generation tasks.
-
$\lambda$-PSD: Scalable Approximate SNR-Optimised Polynomial Stein Discrepancies
λ-PSD reformulates polynomial Stein discrepancies as an SNR² maximization problem via a Rayleigh quotient and supplies an approximate scalable implementation that stabilizes SNR under Gaussian assumptions while improv...
-
A Flat Connection: The Pooling Factor and the Geometry of Centring in Hierarchical MCMC
The Ehresmann connection induced by the Fisher metric on the hierarchical parameter fiber bundle is flat for any smooth posterior, identifying the mixing obstruction as the prior fraction (pooling factor) rather than ...
-
Hessian-informed, Coordinate Friendly Hamiltonian Monte Carlo in Linear Time
A graph-manipulation technique reduces the cost of diagonal-preconditioned RHMC fixed-point iterations from quadratic to linear in dimension for coordinate-friendly targets.
-
Transfer Learning of Multiobjective Indirect Low-Thrust Trajectories Using Diffusion Models and Markov Chain Monte Carlo
A homotopy-plus-MCMC data-generation pipeline trains a mass-conditioned diffusion model that yields 40% more feasible initial costates and a better Pareto front for multiobjective indirect low-thrust transfers than ad...
-
Bayesian Doppler Imaging: Simultaneous Inference of Surface Maps and Geometric Parameters
A fully Bayesian pixel-based Doppler imaging framework uses Gaussian Process priors and Hamiltonian Monte Carlo to simultaneously infer surface maps and geometric parameters from spectral data.
-
Adaptive Meta-Learning Stochastic Gradient Hamiltonian Monte Carlo Simulation for Bayesian Updating of Structural Dynamic Models
AM-SGHMC combines adaptive neural networks with SGHMC to produce a reusable MCMC sampler for Bayesian updating of similar structural dynamic models without per-task retraining.
-
Adaptive Riemannian Manifold Hamiltonian Monte Carlo with Hierarchical Metric
An adaptive hierarchical RMHMC sampler with closed-form leapfrog integrator and automatic mass matrix tuning for efficient MCMC in high-dimensional Bayesian problems.
-
Many Wrongs Make a Right: Leveraging Biased Simulations Towards Unbiased Parameter Inference
Template-Adapted Mixture Model uses many biased simulations for data-driven estimates of signal and background distributions, yielding unbiased signal fraction estimates with well-calibrated uncertainties.
-
Mortality Forecasting as a Flow Field in Tucker Decomposition Space
Mortality forecasting is recast as integrating a flow field through the low-dimensional Tucker decomposition score space of the Human Mortality Database, yielding lower bias and error than Lee-Carter, Hyndman-Ullah, o...
-
AMIGO: a Data-Driven Calibration of the JWST Interferometer
AMIGO is an end-to-end differentiable forward model of JWST AMI that corrects detector systematics to recover high-precision astrometry and detect close high-contrast companions.
-
Bayesian Multivariate Sparse Functional Principal Components Analysis
MSFAST extends the FAST FPCA method to multivariate sparse data via Bayesian modeling with orthonormal splines, standardization, Procrustes alignment, and efficient computation, yielding valid inferences especially in...
-
Alleviating prior dependencies for DESI DR1 clustering fits through reparameterization
Jeffreys prior over EFTofLSS coefficients mitigates projection effects in DESI DR1 power spectrum multipole fits, recentering posteriors for late-time expansion parameters.
-
Continuous-Time Bayesian Networks with Structured Shrinkage Priors for Modelling Multimorbidity Trajectories in Large-Scale Electronic Health Records
A CTBN framework with order-dependent shrinkage priors is developed, compared via simulations, and applied to UK Biobank data to recover cardiometabolic and inflammatory multimorbidity modules.
-
A Bayesian latent Gaussian process framework for aerodynamic uncertainty quantification
A Bayesian latent GP calibration framework for aerodynamic surrogates marginalizes input uncertainty and matches output uncertainty statistics, achieving 94.2-95.8% coverage of true 95% intervals.
-
HIcosmo: a differentiable JAX-based framework for cosmology inference
HIcosmo is a new JAX-based differentiable framework for background cosmology inference that matches Cobaya results while delivering 8.7x CPU and up to 20x GPU speedups.
-
Hierarchical Bayes meets hierarchical forecasting: A flexible framework for level-focused forecasts
A Bayesian hierarchical model integrates coherence penalization and level-specific focus into forecasting estimation, yielding improved predictive accuracy on simulated and Australian tourism data.
-
Optimal sequential two-stage Bayes Factor Design for two-arm clinical Phase II Trials with binary Endpoints
Derives exact operating characteristic corrections and a numerical search over sample sizes to obtain optimal two-stage Bayes factor designs for two-arm binary-endpoint phase II trials that minimize expected sample si...
-
StressDream: Steering Video World Models for Robust Policy Evaluation and Improvement
StressDream optimizes initial noise in diffusion video world models using VLM semantic and plausibility objectives to steer generations toward specified high-impact outcomes for improved policy evaluation.
-
Scalable High-Dimensional Bayesian Field Reconstruction with Finite Elements: Application to 3D Porous Media Flow
A finite-element variational inference method delivers full-covariance Bayesian field reconstruction at dimensions exceeding 400,000 for 3D porous media flow using sparse precision parameterization from SPDE priors.
-
Adaptive Generalized Elliptical Slice Sampling
AGESS combines elliptical slice sampling with diminishing adaptation to self-correct to fast mixing while preserving ergodicity for non-elliptical, multi-modal, and high-dimensional targets.
-
Adaptive Generalized Elliptical Slice Sampling
Proposes an adaptive generalized elliptical slice sampling algorithm that improves efficiency on non-elliptical, non-differentiable, multi-modal and high-dimensional targets and proves ergodicity under general regular...
-
Inferring infectiousness: a joint model of the within-host viral kinetics of SARS-CoV-2
Bayesian joint model infers infectious virus shedding trajectories and derived infectiousness metrics from PCR and other proxies in SARS-CoV-2 using data from five cohorts of roughly 2000 infections.
-
Self-Supervised Laplace Approximation for Bayesian Uncertainty Quantification
SSLA approximates the posterior predictive distribution by refitting Bayesian models on self-predicted data, providing a sampling-free method that improves predictive calibration over classical Laplace approximations ...
-
Self-Supervised Laplace Approximation for Bayesian Uncertainty Quantification
SSLA approximates the posterior predictive distribution by refitting Bayesian models on self-predicted data, yielding a deterministic uncertainty measure that outperforms standard Laplace approximations in calibration...
-
Amortized Variational Inference for Joint Posterior and Predictive Distributions in Bayesian Uncertainty Quantification
An amortized variational framework jointly targets the posterior and posterior-predictive distributions via a KL upper bound and moment regularization, yielding more accurate predictions at lower online cost than two-...
-
Bayesian X-Learner: Calibrated Posterior Inference for Heterogeneous Treatment Effects under Heavy-Tailed Outcomes
Bayesian X-Learner delivers calibrated posterior inference for CATE by combining cross-fitted doubly robust pseudo-outcomes with a Welsch redescending pseudo-likelihood and MCMC sampling.
-
Tempered Sequential Monte Carlo for Trajectory and Policy Optimization with Differentiable Dynamics
Tempered sequential Monte Carlo samples efficiently from a temperature-annealed distribution over controller parameters to solve trajectory and policy optimization under differentiable dynamics.
-
Tempered Sequential Monte Carlo for Trajectory and Policy Optimization with Differentiable Dynamics
Tempered sequential Monte Carlo samples from a Boltzmann-tilted distribution over controllers to optimize trajectories and policies under differentiable dynamics.
-
Combining Bayesian and Frequentist Inference for Laboratory-Specific Performance Guarantees in Copy Number Variation Detection
A hybrid method models squared losses from Bayesian CNV posteriors with a Gamma distribution on validation samples to produce tolerance intervals with valid frequentist coverage, achieving single-digit mean absolute c...
-
Jeffreys Flow: Robust Boltzmann Generators for Rare Event Sampling via Parallel Tempering Distillation
Jeffreys Flow distills Parallel Tempering trajectories via Jeffreys divergence to produce robust Boltzmann generators that suppress mode collapse and correct sampling inaccuracies for rare event sampling.
-
A unified harmonic framework for dark siren cosmology
The GW-galaxy cross-correlation method, unified with spectral sirens in a harmonic framework, can measure H0 to 1% and Omega_m to 5% precision with 2 years of data from next-generation detectors like Einstein Telescop...
-
The Geometric Reasoner: Manifold-Informed Latent Foresight Search for Long-Context Reasoning
TGR performs manifold-informed latent foresight search to boost trajectory coverage in long-context reasoning tasks by up to 13 AUC points with minimal overhead.
-
Stochastic gravitational-wave background search using data from five pulsar timing arrays
Combined five-PTA dataset yields posterior on SGWB power-law amplitude and index consistent with nonzero signal but below 5-sigma significance, with reconstructed angular correlations matching the Hellings-Downs prediction.
-
RefineStat: Efficient Exploration for Probabilistic Program Synthesis
RefineStat improves small language model performance on probabilistic program synthesis by adding semantic constraint enforcement and diagnostic-aware refinement, producing syntactically and statistically reliable cod...
-
Bayesian Neural Network Surrogates for Bayesian Optimization of Carbon Capture and Storage Operations
Comparative evaluation of Bayesian Neural Network surrogates versus Gaussian Processes in Bayesian Optimization applied to Carbon Capture and Storage operations, presented as the first such application in reservoir en...
-
On Divergence Measures for Training GFlowNets
Introduces statistically efficient estimators for Renyi-α, Tsallis-α, reverse and forward KL divergences with REINFORCE and score-matching control variates for faster GFlowNet training.
-
dynesty: A Dynamic Nested Sampling Package for Estimating Bayesian Posteriors and Evidences
dynesty is an open-source Python package for dynamic nested sampling that improves efficiency in Bayesian posterior and evidence estimation compared to MCMC on certain problems.
-
Measurement Induced Confounding
Measurement error in latent confounders propagates bias into ATE estimates under conventional adjustment methods, which joint Bayesian estimation of measurement, treatment, and outcome models can correct.
-
Posterior sampling in the Age of Emulators
Compares MH, MALA, HMC, NUTS, and AIES on differentiable likelihood emulators for ΛCDM and sterile-neutrino models, finding MALA and MH competitive in wall time despite NUTS needing fewer samples.
-
Constrained Weighted Bayesian Bootstrap
Constrained weighted Bayesian bootstrap extends weighted Bayesian bootstrap to constrained posteriors with asymptotics matching restricted MLE and is demonstrated on option pricing.
-
Bayesian reversal of the liquid level trajectory in a draining tank for pollution forensics
Bayesian inversion infers initial liquid level in a draining tank from final level and duration via Torricelli's law augmented by an empirical discrepancy function, validated on water tank experiments.
-
StanBKT: Rethinking Parameter Estimation in Bayesian Knowledge Tracing
StanBKT provides a unified Bayesian inference framework for BKT models supporting HMC, variational inference, and hierarchical variants, evaluated on ASSISTments and intervention datasets.
-
A Scalable Parametric Item Calibration Engine (SPICE) for Explanatory IRT with Sparse Data
SPICE is a scalable Bayesian MCMC engine for explanatory IRT calibration on sparsely linked persons and items in large assessment banks.
-
The impact of observation density on Bayesian inversion of latent dynamics in shock-dominated flows
A latent-space reduced-order model using autoencoders and learned dynamics enables Bayesian recovery of initial density and pressure in Sod shock tube simulations, with posterior uncertainty contracting substantially ...
-
A Cubing Strategy for Identifying Stable Hyperparameter Regions for Uncertainty Quantification in Spatial Deep Learning
A recursive cubing framework identifies stable hyperparameter regions for MC dropout uncertainty quantification in spatial deep learning and produces competitive or superior predictive intervals versus a statistical b...
-
Uncertainty-Aware Prediction of Lung Tumor Growth from Sparse Longitudinal CT Data via Bayesian Physics-Informed Neural Networks
Bayesian PINN integrates Gompertz dynamics and HMC sampling to predict tumor growth from sparse CT data, achieving log-space RMSE of 0.20 with well-calibrated 95% credible intervals on 30 NLST patients.
-
Uncertainty Quantification for Cardiac Shape Reconstruction with Deep Signed Distance Functions via MCMC methods
Deep signed distance functions combined with MCMC sampling enable uncertainty-aware reconstruction of left and right ventricles from limited data.
-
Singularity Formation: Synergy in Theoretical, Numerical and Machine Learning Approaches
The work introduces a modulation-based analytical method for singularity proofs in singular PDEs and refines ML techniques like PINNs and KANs to identify blowup solutions, with application to the open 3D Keller-Segel...
-
LITMUS: Bayesian Lag Recovery in Reverberation Mapping with Fast Differentiable Models
LITMUS introduces a differentiable Bayesian lag recovery framework that outperforms JAVELIN on OzDES-like mock data by reducing false positives from seasonal aliasing.
-
Temporal Point Process Modeling of Aggressive Behavior Onset in Psychiatric Inpatient Youths with Autism
Applies self-exciting temporal point processes to model clustered aggression onsets in inpatient autistic youth and reports better fit than Poisson baselines.
-
Measurement Induced Confounding
Measurement error in latent confounders produces biased ATE estimates and miscalibrated intervals under conventional adjustment; a Bayesian joint model of measurement, treatment, and outcome is proposed to correct it.
-
Stochastic Volatility in Mean Models with Heavy Tails: A Fast Approximate Bayesian Inference Using Hidden Markov Models
Extends SVM models with SMN heavy tails via HMM representation for fast approximate Bayesian inference, claiming ~10x speedup over MCMC with maintained accuracy.
-
Pointwise is Pointless? A Multimodal Ablation Study for Precipitation Nowcasting with Graph Neural Networks
A multimodal GNN ablation for Nordic precipitation nowcasting shows sparse point observations improve station and onset scores while NWP and CRPS losses improve radar-grid performance, indicating local and field skill...
-
Revisiting the Exo-Mercury Candidate GJ 367 b with ESPRESSO and a Self-Consistent Tidal Distortion Model
Revised mass of 0.503 M_Earth and radius of 0.736 R_Earth for GJ 367 b give a density of 6.9 g cm^{-3} and an iron fraction of 50-70% via new tidal and composition modeling.
-
Uncertainty in Physics and AI: Taxonomy, Quantification, and Validation
A unified taxonomy of uncertainty in ML for physics is introduced together with validation tools such as coverage, calibration, and proper scoring rules, illustrated on regression and classification tasks.
-
Bayesian Inference of Nonlinear Malaria Dynamics in Ghana via an Ensemble Markov Chain Monte Carlo Sampler
Bayesian ensemble MCMC fits a cubic-plus-damped-oscillatory model to Ghana malaria data, reports R² > 0.995, and extrapolates rising case counts through 2026 with widening uncertainty.
-
Methodological Frontiers in 21-cm Intensity Mapping: the Treatment of Systematics and Foreground Contamination
A review chapter summarizing advances in map-making and component-separation techniques for 21-cm intensity mapping to address systematics and foreground contamination in simulations for SKA-Mid.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.