PulseAugur
EN
LIVE 09:57:27

New Legendre Jumper Martingales enhance online exchangeability testing

Researchers have introduced a new family of conformal test martingales called Legendre Jumper Martingales, designed to detect deviations from exchangeability in data. These martingales extend the Simple Jumper martingale by using polynomial betting functions of arbitrary degrees to identify variance, skewness, and higher-order deviations. The Variational Legendre Jumper addresses the state space expansion issue of the Product Legendre Jumper by using a mean-field approximation for linear time complexity. Empirical results on a real-world classification task show that these combined methods outperform single-degree martingales, with the composite variant being recommended for unknown shift timescales. AI

RANK_REASON The cluster contains a research paper detailing a new statistical method. [lever_c_demoted from research: ic=1 ai=0.4]

Read on arXiv stat.ML →

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

New Legendre Jumper Martingales enhance online exchangeability testing

COVERAGE [1]

  1. arXiv stat.ML TIER_1 English(EN) · Johan Hallberg Szabadváry ·

    Betting on Moments: Legendre Jumper Martingales for Online Exchangeability Testing

    We present a family of conformal test martingales based on shifted Legendre polynomials, which extends the Simple Jumper martingale. The Simple Legendre Jumper substitutes the linear betting function with a polynomial of arbitrary degree, thereby facilitating the detection of var…