FAIR Universe HiggsML Uncertainty Challenge Competition

2024-10-03

arXiv.org

AI Summary

Systematic (epistemic) uncertainties arising from modeling deficiencies in high-energy physics simulators remain difficult to quantify rigorously. Method: This project introduces a novel paradigm integrating physics-informed priors with bias-aware machine learning. We construct the first large-scale AI competition platform dedicated to systematic uncertainty quantification, incorporating Monte Carlo simulation, surrogate modeling, and uncertainty calibration techniques to enable robust parameter inference on biased simulation data. Contributions/Results: We release the first standardized uncertainty benchmark dataset for Higgs physics; develop an interpretable, calibratable framework for systematic error assessment; and significantly improve model robustness against simulator misspecification and accuracy of uncertainty estimation. These advances establish a trustworthy AI pathway for precision measurements in particle physics.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Probabilistic Programming
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The FAIR Universe -- HiggsML Uncertainty Challenge focuses on measuring the physics properties of elementary particles with imperfect simulators due to differences in modelling systematic errors. Additionally, the challenge is leveraging a large-compute-scale AI platform for sharing datasets, training models, and hosting machine learning competitions. Our challenge brings together the physics and machine learning communities to advance our understanding and methodologies in handling systematic (epistemic) uncertainties within AI techniques.

Problem

• Measuring elementary particle properties with imperfect simulators
• Handling systematic epistemic uncertainties in AI techniques
• Leveraging large-scale AI platforms for physics challenges

Innovation

• Leveraging AI platform for competitions
• Handling systematic uncertainties with AI
• Advancing methodologies for imperfect simulators