{"id":"arxiv/2606.20859","url":"https://arxiv.org/abs/2606.20859","tags":["stat.ML","cs.LG","math.ST"],"score":0,"title":"Betting on Moments: Legendre Jumper Martingales for Online Exchangeability Testing","domain":"stat.ML","source":"arxiv","authors":["Johan Hallberg Szabadváry"],"reviews":0,"subject":"stat.ML","summary":"A fundamental assumption in statistics and machine learning is that ``the future looks like the past,'' formalized as exchangeability: the joint data distribution is order-invariant. In practice, this assumption is often violated due to distribution shifts over time. Early detection of exchangeability violations is crucial to prevent performance degradation and enable timely interventions like model retraining. Conformal test martingales offer a flexible, distribution-free framework for sequential exchangeability testing with guaranteed false-alarm rate control by betting against the…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-06-18","reactions":{}}