• Corpus ID: 248693535

Bias and Priors in Machine Learning Calibrations for High Energy Physics

@inproceedings{Gambhir2022BiasAP,
  title={Bias and Priors in Machine Learning Calibrations for High Energy Physics},
  author={Rikab Gambhir and Benjamin Philip Nachman and Jesse Thaler},
  year={2022}
}
Machine learning offers an exciting opportunity to improve the calibration of nearly all reconstructed objects in high-energy physics detectors. However, machine learning approaches often depend on the spectra of examples used during training, an issue known as prior dependence. This is an undesirable property of a calibration, which needs to be applicable in a variety of environments. The purpose of this paper is to explicitly highlight the prior dependence of some machine learning-based… 

Figures from this paper

Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics
TLDR
This paper presents a machine learning framework for performing frequentist maximum likelihood inference with Gaussian uncertainty estimation, which also quantifies the mutual information between the unobservable and measured quantities.

References

SHOWING 1-10 OF 124 REFERENCES
Uncertainty Aware Learning for High Energy Physics With A Cautionary Tale
TLDR
This work studies the use of a classifier that is fully aware of uncertainties and their corresponding nuisance parameters, and shows that this dependence can actually enhance the sensitivity to parameters of interest compared to baseline approaches.
Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics
TLDR
This paper presents a machine learning framework for performing frequentist maximum likelihood inference with Gaussian uncertainty estimation, which also quantifies the mutual information between the unobservable and measured quantities.
Adversarially-trained autoencoders for robust unsupervised new physics searches
TLDR
It is proposed to combine the autoencoder with an adversarial neural network to remove its sensitivity to the smearing of the final-state objects and it is shown that one can achieve a robust anomaly detection in resonance-induced resonance.
Parametrizing the detector response with neural networks
In high energy physics, characterizing the response of a detector to radiation is one of the most important and fundamental experimental tasks. In many cases, this task is accomplished by
Enhancing searches for resonances with machine learning and moment decomposition
TLDR
A new set of tools using a novel moment loss function (Moment Decomposition or MoDe) which relax the assumption of independence without creating structures in the background to enhance the sensitivity to new physics without compromising the fidelity of the background estimation.
Automating the ABCD method with machine learning
TLDR
Automating the ABCD method with machine learning significantly improves performance in terms of ABCD closure, background rejection and signal contamination, and uncovers a previously unappreciated aspect of the method: its accuracy hinges on having low signal contamination in control regions not just overall, but relative to the signal fraction in the signal region.
Calorimetric Measurement of Multi-TeV Muons via Deep Regression
The performance demands of future particle-physics experiments investigating the high-energy frontier pose a number of new challenges, forcing us to find improved solutions for the detection,
Machine learning uncertainties with adversarial neural networks
TLDR
Using adversarial networks, a priori known sources of systematic and theoretical uncertainties during the training pave the way to a more reliable event classification on an event-by-event basis, as well as novel approaches to perform parameter fits of particle physics data.
Deep-learning jets with uncertainties and more
TLDR
The main features of Bayesian versions of established deep-learning taggers are illustrated and it is shown how they capture statistical uncertainties from finite training samples, systematics related to the jet energy scale, and stability issues through pile-up.
A convolutional neural network based cascade reconstruction for the IceCube Neutrino Observatory
TLDR
A reconstruction method based on convolutional architectures and hexagonally shaped kernels is presented that can improve upon the reconstruction accuracy, while reducing the time necessary to run the reconstruction by two to three orders of magnitude.
...
...