Solomonoff induction violates Nicod's criterion

DOI10.1007/978-3-319-24486-0_23MaRDI QIDQ2835642zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Jan Leike, Marcus Hutter

Publication date 30 November 2016

Published in Lecture Notes in Computer Science (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1507.04121

zbMATH Keywords

confirmationequivalence conditionBayesian reasoningdisconfirmationHempel's paradoxSolomonoff normalization

Mathematics Subject Classification ID

Probability and inductive logic (03B48) Algorithmic information theory (Kolmogorov complexity, etc.) (68Q30)

Abstract: Nicod's criterion states that observing a black raven is evidence for the hypothesis H that all ravens are black. We show that Solomonoff induction does not satisfy Nicod's criterion: there are time steps in which observing black ravens decreases the belief in H. Moreover, while observing any computable infinite string compatible with H, the belief in H decreases infinitely often when using the unnormalized Solomonoff prior, but only finitely often when using the normalized Solomonoff prior. We argue that the fault is not with Solomonoff induction; instead we should reject Nicod's criterion.

Recommendations

• Natural descriptions and anthropic bias: extant problems in Solomonoff induction
• Modes of convergence to the truth: steps toward a better epistemology of induction
• On universal prediction and Bayesian confirmation
• Theory and Applications of Models of Computation
• Principles of Solomonoff Induction and AIXI

Cites work

• scientific article; zbMATH DE number 3489017 (Why is no real title available?)
• (Non-)Equivalence of Universal Priors
• A formal theory of inductive inference. Part I
• A philosophical treatise of universal induction
• An introduction to Kolmogorov complexity and its applications
• Complexity-based induction systems: Comparisons and convergence theorems
• Hempel's Raven Paradox: A Lacuna in the Standard Bayesian Solution
• I.—STUDIES IN THE LOGIC OF CONFIRMATION (I.)
• Merging of Opinions with Increasing Information
• New error bounds for Solomonoff prediction
• On the relation between descriptional complexity and algorithmic probability
• On universal prediction and Bayesian confirmation
• Probability Theory
• The generalized radially symmetric wave equation

This page was built for publication: Solomonoff induction violates Nicod's criterion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2835642)