Algorithmic Identification of Probabilities Is Hard - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Algorithmic Identification of Probabilities Is Hard

Résumé

Suppose that we are given an infinite binary sequence which is random for a Bernoulli measure of parameter p. By the law of large numbers, the frequency of zeros in the sequence tends to p, and thus we can get better and better approximations of p as we read the sequence. We study in this paper a similar question, but from the viewpoint of inductive inference. We suppose now that p is a computable real, and one asks for more: as we are reading more and more bits of our random sequence, we have to eventually guess the exact parameter p (in the form of its Turing code). Can one do such a thing uniformly for all sequences that are random for computable Bernoulli measures, or even for a 'large enough' fraction of them? In this paper, we give a negative answer to this question. In fact, we prove a very general negative result which extends far beyond the class of Bernoulli measures.
Fichier principal
Vignette du fichier
paper_learning_measures_remasterized3.pdf (279.79 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01397246 , version 1 (17-11-2016)

Identifiants

Citer

Laurent Bienvenu, Benoit Monin, Alexander Shen. Algorithmic Identification of Probabilities Is Hard. ALT: Algorithmic Learning Theory, Oct 2014, Bled, Slovenia. pp.85-95, ⟨10.1007/978-3-319-11662-4_7⟩. ⟨hal-01397246⟩
334 Consultations
80 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More