Implicit Maximum Likelihood Estimation
This addresses a fundamental challenge in machine learning for researchers and practitioners working with implicit models, though it appears incremental as it builds on existing likelihood-based frameworks.
The paper tackles the problem of estimating parameters in implicit probabilistic models, which lack explicit likelihood functions, by developing a method equivalent to maximum likelihood estimation under certain conditions and demonstrating encouraging experimental results.
Implicit probabilistic models are models defined naturally in terms of a sampling procedure and often induces a likelihood function that cannot be expressed explicitly. We develop a simple method for estimating parameters in implicit models that does not require knowledge of the form of the likelihood function or any derived quantities, but can be shown to be equivalent to maximizing likelihood under some conditions. Our result holds in the non-asymptotic parametric setting, where both the capacity of the model and the number of data examples are finite. We also demonstrate encouraging experimental results.