New research explores stochastic extragradient methods for variational inequalities

Researchers have published a new paper detailing advancements in stochastic extragradient (SEG) methods for solving monotone variational inequality problems. The study addresses gaps in understanding the behavior of both same-sample (S-SEG) and independent-sample (I-SEG) variants, particularly in scenarios with unbounded domains or varying sample Lipschitz parameters. The findings indicate that S-SEG can be sensitive to samplewise Lipschitz parameters, and that standard assumptions like mean Lipschitzness or bounded variance are insufficient for convergence in all cases. The paper also establishes high-probability restricted-gap convergence for both SEG variants under relaxed conditions and demonstrates that certain step-size strategies guaranteeing convergence for I-SEG may fail for S-SEG. AI

IMPACT Advances theoretical understanding of optimization algorithms relevant to machine learning.

RANK_REASON The cluster contains a single academic paper published on arXiv detailing new theoretical findings in optimization algorithms. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• alphaXiv
• CatalyzeX Code Finder for Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Independent-Sample Stochastic Extragradient
• Influence Flower
• Same-Sample Stochastic Extragradient
• ScienceCast
• Stochastic Extragradient

• paper
• other

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