Sequential Neural Methods for Likelihood-free Inference
By Conor Durkan, George Papamakarios, Iain Murray. 2018
26 citations recorded by the source corpus.
Likelihood-free inference refers to inference when a likelihood function cannot be explicitly evaluated, which is often the case for models based on simulators. Most of the literature is based on sample-based `Approximate Bayesian Computation' methods, but recent work suggests that approaches based on deep neural conditional density estimators can obtain state-of-the-art results with fewer simulations. The neural approaches vary in how they choose which simulations to run and what they learn: an approximate posterior or a surrogate likelihood. This work provides some direct controlled comparisons between these choices.
Subfields: Artificial Intelligence, Statistics and Probability, Probabilistic & Bayesian Methods
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Benjamin Dan Wandelt
- Justin Alsing
- S. Feeney
- David S. Greenberg
- Marcel Nonnenmacher
- Jakob H. Macke
- Tom Charnock
- Gilles Louppe
In Statistics and Probability
- Benjamin Dan Wandelt
- Justin Alsing
- S. Feeney
- David S. Greenberg
- Marcel Nonnenmacher
- Jakob H. Macke
- Tom Charnock
- Gilles Louppe
In Probabilistic & Bayesian Methods
- Benjamin Dan Wandelt
- Justin Alsing
- David S. Greenberg
- Marcel Nonnenmacher
- Jakob H. Macke
- S. Feeney
- Tom Charnock
- Gilles Louppe
Citations in the corpus
- Automatic Posterior Transformation for Likelihood-Free Inference (2019, methodology): “(Papamakarios et al., 2018) quantitatively compared these algorithms with classical ABC methods (SMC-ABC and SL).”
- Fast likelihood-free cosmology with neural density estimators and active learning (2019, result): “However, Durkan et al. (2018) reported that the two approaches gave broadly similar performance in a number of experiments in the context of DELFI.”