Get Started
Research questionHow can SGMCMC hyperparameters be tuned without Metropolis-Hastings acceptance rates when computation is limited?SGMCMC performance depends on choices such as step size, mini-batch size, and leapfrog steps. These methods often lack the Metropolis-Hastings acceptance rate used by standard tuning procedures, while limited computation makes evaluating many candidate settings difficult.
Machine Learning
Research Paper
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.HyperMC: Multi-Fidelity Hyperparameter Tuning for Stochastic Gradient MCMCThe source concerns stochastic gradient MCMC hyperparameter tuning using kernel Stein discrepancy evaluations and multi-fidelity resource allocation. Evidence covers logistic regression, probabilistic matrix factorization, and Bayesian neural networks, with outcomes including posterior approximation and predictive calibration; theoretical guarantees rely on approximation and concentration conditions for estimated KSD.research paper · Sep 2, 2026
Related questions
How can MCMC explore multimodal targets while preserving mode weights and avoiding approximation bias?How can production systems tune inference-time configurations online from noisy live feedback without representative validation data?How should Bayesian optimization allocate replications for expensive, high-variance stochastic objectives?How can active preference learning obtain scalable, calibrated uncertainty for neural reward models without full Bayesian inference?
Home
Topics
Search
Library