2025/03/04 by Vladimir Vovk, Vovk, Vladimir · 2 voices
Computer Science · Mathematics · #68Q32 (Primary) 62G15 #68T05 (Secondary) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Methodology (stat.ME) #cs.LG #stat.ME
paper · pdf · doi:10.48550/arxiv.2503.02803
arxiv published 2025/03/04 · arxiv updated 2025/07/07
This paper introduces inductive randomness predictors, which form a proper superset of inductive conformal predictors but have the same principal property of validity under the assumption of randomness (i.e., of IID data). It turns out that every non-trivial inductive conformal predictor is strictly dominated by an inductive randomness predictor, although the improvement is not great, at most a factor of e≈2.72 in the case of e-prediction. The dominating inductive randomness predictors are more complicated and more difficult to compute; besides, an improvement by a factor of e is rare. Therefore, this paper does not suggest replacing inductive conformal predictors by inductive randomness predictors and only calls for a more detailed study of the latter.