Approximate Bayesian Computation applied to Metrology for Nuclear Safeguards Journal of Physics: Conference Series

Abstract

Approximate Bayesian Computation (ABC) is an inference option if a likelihood for the measurement data is not available, but a forward model is available that outputs predicted observables for specified input parameters. This paper applies ABC to the enrichment meter principle (EMP) method, which is used in nuclear safeguards and does not have an explicit likelihood. A key aspect of metrology is uncertainty quantification (UQ), approached from physical first principles (“bottom-up”) or approached empirically by comparing measurements from different methods and/or laboratories (“top-down”). Although ABC is not commonly used in metrology, the EMP example illustrates advantages in ABC compared to current bottom-up approaches. ABC is also shown to be useful in top-down UQ. As a diagnostic, in bottom-up and top-down applications of ABC, the actual coverages of probability intervals are compared to the true coverages. If an ABC-based interval for a parameter is constructed to contain approximately 95% of the true parameter values, then it is important to verify that the actual coverage is close to 95%. It is shown that one advantage of ABC compared to other Bayesian approaches is its apparent robustness to miss-specifying the likelihood while maintaining good agreement between nominal and actual coverage.