{"id":"arxiv/2611.11188","url":"https://arxiv.org/abs/2611.11188","tags":["stat.ME","math.PR","math.ST"],"score":0,"title":"Heavy tails in Gaussian-process surrogates: process-level profiling degeneracy versus observation-level identifiability","domain":"stat.ME","source":"arxiv","authors":["Neelesh S. Upadhye","Rohit Kumar Mani"],"reviews":0,"subject":"stat.ME","summary":"In this article, we study heavy-tailed alternatives to Gaussian-process surrogates, namely Student-$t$ processes, $α$-stable processes and Gaussian processes with heavy-tailed noise. These are Gaussian scale mixtures, and the position of the mixing variable decides whether the tails can be learned from a single realization. For process-level mixtures with a free amplitude, the profile likelihood of the mixing law is data-free and is maximized by the Gaussian law, the plug-in predictive is a data-free transformation of the Gaussian-process predictive, and no estimator can be consistent for…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-11-08","reactions":{}}