Markov Chains and Monte Carlo Methods

abcneuralnet: Simulation-based Inference with Bayesian Neural Networks and Explainable AI for Population Genomics Inference

abcneuralnet provides a Bayesian Neural Network framework with explainable AI for population genomics inference, offering calibrated prediction intervals and uncertainty decomposition.

Efficient approximate Bayesian inference for large, heterogeneous biological datasets

The paper introduces an adaptive minibatch approach for approximate Bayesian inference, improving speed-precision tradeoffs for large, heterogeneous biological datasets.

BayesForge: A Bayesian Inference library for Python, R, and Julia

BayesForge provides a unified Bayesian inference framework across Python, R, and Julia, enabling high-performance computation with hardware acceleration and demonstrating significant speedups over existing methods.

Observation-masked neural posterior estimation for heterogeneous epidemiological surveillance data

The paper introduces OM-NPE, a Bayesian inference framework for heterogeneous epidemiological data, demonstrating its ability to recover parameters and uncertainties efficiently.

Overinflation and overconcentration: why Cauchy perturbation kernels are the right choice for ABC-SMC

This paper shows that Cauchy perturbation kernels outperform Normal kernels in ABC-SMC by addressing overinflation and overconcentration issues, improving performance in high dimensions.

Efficient Bayesian inference for ordinary differential equation models from experimental data with uncertain measurement times

This paper presents a Bayesian framework for parameter inference in ODE models that accounts for uncertain measurement times, improving reliability and efficiency in dynamic modeling.

Composite Certainty: Addressing Metric Degeneracy in Parameter Inference for Model-Based Diagnostics

The paper introduces a Composite Certainty Framework to address metric degeneracy in parameter inference by aggregating multiple uncertainty metrics into a single consensus score, revealing parameter-specific identifiability profiles and clinical thresholds.

Trajectory matching-random walk feasibility sampling for simulating heterogeneous dynamics in mechanistic models

This paper introduces TM-RWFS, a method to generate heterogeneity in mechanistic models by constraining trajectories between data bounds, showing it efficiently reproduces biological variability.

ABC for high-dimensional modular models via MCMC samples

The paper introduces an improved ABC-PMC method for high-dimensional modular models by leveraging precomputed MCMC samples and adaptive summary statistics, significantly reducing computational cost.

CHRRT: boosting coordinate hit-and-run with rounding by thinning

The paper introduces CHRRT, a modified Coordinate Hit-and-Run with Rounding algorithm that uses thinning to improve computational efficiency for sampling convex polytopes, achieving up to three orders of magnitude faster effective sample size per time.

Algorithmic Parameter Estimation and Uncertainty Quantification for Hodgkin-Huxley Neuron Models

The paper introduces a Bayesian MCMC method with adaptive parallel tempering to estimate parameters and quantify uncertainty in Hodgkin-Huxley neuron models using voltage trace data.

Robust adaptive distance functions for approximate Bayesian inference on outlier-corrupted data

This paper introduces robust distance functions for ABC that handle outliers, showing improved performance on corrupted data while maintaining comparable results on clean data.

No-U-Turn sampling for phylogenetic trees

This paper introduces a new Bayesian inference algorithm for phylogenetic trees using No-U-Turn sampling, which converges faster and explores the posterior distribution more efficiently than traditional methods.

Tunings for leapfrog integration of Hamiltonian Monte Carlo for estimating genetic parameters

This study applies Hamiltonian Monte Carlo to animal breeding data, identifies optimal leapfrog integration parameters, and shows superior performance over Gibbs sampling in models using pedigree or genomic information.

Multi-level Approximate Bayesian Computation

The paper introduces a multi-level approximate Bayesian computation framework for inference in mechanistic models of biological systems.

Thermodynamic integration for dynamic causal models

This paper introduces thermodynamic integration for dynamic causal modeling, showing it provides more robust log evidence estimates than variational Bayes in complex models.

Inverse sensitivity analysis of mathematical models avoiding the curse of dimensionality

The paper introduces a novel Markov Chain Monte Carlo method for inverse sensitivity analysis of deterministic biological models, demonstrating its application to several mathematical models.