Robust semi-parametric signal detection in particle physics with classifiers decorrelated via optimal transport

2024-09-10

AI Summary

In particle physics signal searches, systematic uncertainties in background modeling bias supervised classifiers, degrading the statistical power and robustness of bump-hunt analyses. To address this, we propose an optimal transport (OT)-based classifier decorrelation method—the first to leverage OT for constructing classifier outputs statistically independent of protected variables (e.g., background systematic parameters)—and establish a theoretical link between decorrelation and detection power. Integrated with a semiparametric mixture model, our approach enables robust signal enrichment on real data. Experiments demonstrate that, under moderate background misspecification, our method significantly outperforms adversarial training and reweighting baselines: it achieves higher detection power, ensures controllable and interpretable statistical significance, and maintains robustness without sacrificing sensitivity.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Adversarial Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Searches of new signals in particle physics are usually done by training a supervised classifier to separate a signal model from the known Standard Model physics (also called the background model). However, even when the signal model is correct, systematic errors in the background model can influence supervised classifiers and might adversely affect the signal detection procedure. To tackle this problem, one approach is to use the (possibly misspecified) classifier only to perform a preliminary signal-enrichment step and then to carry out a bump hunt on the signal-rich sample using only the real experimental data. For this procedure to work, we need a classifier constrained to be decorrelated with one or more protected variables used for the signal detection step. We do this by considering an optimal transport map of the classifier output that makes it independent of the protected variable(s) for the background. We then fit a semi-parametric mixture model to the distribution of the protected variable after making cuts on the transformed classifier to detect the presence of a signal. We compare and contrast this decorrelation method with previous approaches, show that the decorrelation procedure is robust to moderate background misspecification, and analyse the power of the signal detection test as a function of the cut on the classifier.

Problem

• Detecting new physics signals while mitigating systematic background model errors
• Developing classifiers decorrelated from protected variables using optimal transport
• Enhancing signal detection robustness through semiparametric mixture modeling

Innovation

• Classifier decorrelated via optimal transport
• Semiparametric mixture model for signal detection
• Robust signal enrichment with background independence