Gaussian Process Accelerated Feldman-Cousins Approach for Physical Parameter Inference
This work addresses a computational bottleneck for physicists conducting exact statistical inference in experiments like neutrino oscillation measurements, representing an incremental improvement.
The paper tackles the computational intensity of the Feldman-Cousins method for statistical inference in high-energy physics by proposing an algorithm using Gaussian processes to accelerate confidence interval construction, achieving speed-ups of 5x in one dimension and 10x in two dimensions with over 99.5% accuracy.
The unified approach of Feldman and Cousins allows for exact statistical inference of small signals that commonly arise in high energy physics. It has gained widespread use, for instance, in measurements of neutrino oscillation parameters in long-baseline experiments. However, the approach relies on the Neyman construction of the classical confidence interval and is computationally intensive as it is typically done in a grid-based fashion over the entire parameter space. In this letter, we propose an efficient algorithm for the Feldman-Cousins approach using Gaussian processes to construct confidence intervals iteratively. We show that in the neutrino oscillation context, one can obtain confidence intervals 5 times faster in one dimension and 10 times faster in two dimensions, while maintaining an accuracy above 99.5%.