Self-supervised learning enhances resonance mass regression in physics

Researchers have developed a self-supervised learning approach using a Transformer encoder pre-trained with VICReg to improve resonance mass regression in high energy physics. This method aims to overcome the limitations of traditional supervised learning, which often struggles with systematic uncertainties and distribution shifts in collider experiments. The pre-trained model demonstrates more stable performance and sharper resonance peaks compared to a supervised model, particularly under realistic corruptions and for heavy resonances in SUSY-like cascade decays. AI

IMPACT Introduces a more robust method for analyzing particle decay data, potentially improving sensitivity in new physics searches.

RANK_REASON Academic paper detailing a novel application of self-supervised learning in high energy physics. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• foundation model
• high energy physics
• supersymmetry
• Transformer encoder
• Vicreg

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →