Aaron Roth's new paper: randomized sample-complexity separation for calibrated multiaccuracy

Aaron Roth, with Lu Jiuyao and Georgy Noarov, released a paper on calibrated multiaccuracy, arguing multicalibration is often overkill and showing a strict polynomial minimax sample-complexity separation between deterministic and randomized predictors for the problem, unlike multicalibration which needs no randomization.

2026-11-03 ~ 2026-11-03 · 4 related posts

• Multicalibration is often overkill: calibrated multiaccuracy needs fewer samples — Aaroth · 2026-11-03
• Sample Complexity of Calibrated Multiaccuracy Depends on Test-Time Randomization — Aaroth · 2026-11-03
• Calibrated multiaccuracy shows sharp randomized-vs-deterministic sample complexity gap — Aaroth · 2026-11-03
• Aaron Roth shares calibrated multiaccuracy paper with Jiuyao Lu and Georgy Noarov — Aaroth · 2026-11-03