Efficient multifidelity likelihood-free Bayesian inference with adaptive computational resource allocation
From MaRDI portal
Abstract: Likelihood-free Bayesian inference algorithms are popular methods for calibrating the parameters of complex, stochastic models, required when the likelihood of the observed data is intractable. These algorithms characteristically rely heavily on repeated model simulations. However, whenever the computational cost of simulation is even moderately expensive, the significant burden incurred by likelihood-free algorithms leaves them unviable in many practical applications. The multifidelity approach has been introduced (originally in the context of approximate Bayesian computation) to reduce the simulation burden of likelihood-free inference without loss of accuracy, by using the information provided by simulating computationally cheap, approximate models in place of the model of interest. The first contribution of this work is to demonstrate that multifidelity techniques can be applied in the general likelihood-free Bayesian inference setting. Analytical results on the optimal allocation of computational resources to simulations at different levels of fidelity are derived, and subsequently implemented practically. We provide an adaptive multifidelity likelihood-free inference algorithm that learns the relationships between models at different fidelities and adapts resource allocation accordingly, and demonstrate that this algorithm produces posterior estimates with near-optimal efficiency.
Recommendations
- Multifidelity approximate Bayesian computation
- Multifidelity approximate Bayesian computation with sequential Monte Carlo parameter sampling
- Multifidelity multilevel Monte Carlo to accelerate approximate Bayesian parameter inference for partially observed stochastic processes
- Likelihood-free inference with deep Gaussian processes
- Likelihood-free MCMC
Cites work
- Adaptive multi-fidelity polynomial chaos approach to Bayesian inference in inverse problems
- Adaptive multiple importance sampling
- An adaptive sequential Monte Carlo method for approximate Bayesian computation
- An adaptive surrogate modeling based on deep neural networks for large-scale Bayesian inverse problems
- Bayesian Synthetic Likelihood
- Delayed Acceptance ABC-SMC
- Estimation of parameters for macroparasite population evolution using approximate Bayesian computation
- Julia: a fresh approach to numerical computing
- Lazy ABC
- MFNets: data efficient all-at-once learning of multifidelity surrogates as directed networks of information sources
- Multifidelity approaches for optimization under uncertainty
- Multifidelity approximate Bayesian computation
- Multifidelity approximate Bayesian computation with sequential Monte Carlo parameter sampling
- Multifidelity multilevel Monte Carlo to accelerate approximate Bayesian parameter inference for partially observed stochastic processes
- Multilevel approximate Bayesian approaches for flows in highly heterogeneous porous media and their applications
- Multilevel Monte Carlo in approximate Bayesian computation
- Multilevel rejection sampling for approximate Bayesian computation
- Optimal model management for multifidelity Monte Carlo estimation
- Rapid Bayesian Inference for Expensive Stochastic Models
- Sequential Monte Carlo without likelihoods
- Survey of multifidelity methods in uncertainty propagation, inference, and optimization
- The frontier of simulation-based inference
- The pseudo-marginal approach for efficient Monte Carlo computations
- Unbiased estimation with square root convergence for SDE models
Cited in
(2)
This page was built for publication: Efficient multifidelity likelihood-free Bayesian inference with adaptive computational resource allocation
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6202141)
