This paper introduces a novel framework connecting betting-based sequential tests with Blackwell approachability through support-function residuals. It establishes an exact pathwise identity relating the distance to a target set with the sum of prediction errors and regret. The research formulates a controlled stochastic experiment where adaptive nulls lead to sublinear regret and exponential wealth under persistent mean separation, with applications to bounded means, kernel MMD, and heterogeneous data sources. AI
IMPACT Introduces theoretical advancements in sequential testing and statistical inference, potentially impacting future AI research in decision-making and hypothesis testing.
RANK_REASON The cluster contains an academic paper published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
