Mechanical Turkeys

Journal of Philosophical Logic:1-22 (forthcoming)
  Copy   BIBTEX

Abstract

Some learning strategies that work well when computational considerations are abstracted away from become severely limiting when such considerations are taken into account. We illustrate this phenomenon for agents who attempt to extrapolate patterns in binary data streams chosen from among a countable family of possibilities. If computational constraints are ignored, then two strategies that will always work are learning by enumeration (enumerate the possibilities---in order of simplicity, say---then search for the one earliest in the ordering that agrees with your data and use it to predict the next data point) and Bayesian learning. But there are many types of computable data streams that, although they can be successfully extrapolated by computable agents, cannot be handled by any computable learner by enumeration. And while there is a sense in which Bayesian learning is a fully general strategy for computable learners, the ability to mimic powerful learners comes at a price for Bayesians: they cannot, in general, become highly confident of their predictions in the limit of large data sets and they cannot, in general, use priors that incorporate all relevant background knowledge.

Other Versions

No versions found

Similar books and articles

Bayesian merging of opinions and algorithmic randomness.Francesca Zaffora Blando - forthcoming - British Journal for the Philosophy of Science.
Classifications of Computable Structures.Karen Lange, Russell Miller & Rebecca M. Steiner - 2018 - Notre Dame Journal of Formal Logic 59 (1):35-59.
Turing computable embeddings.Julia Knight, Sara Miller & M. Vanden Boom - 2007 - Journal of Symbolic Logic 72 (3):901-918.

Analytics

Added to PP
2024-11-29

Downloads
115 (#186,188)

6 months
115 (#50,062)

Historical graph of downloads
How can I increase my downloads?

Author's Profile

Gordon Belot
University of Michigan, Ann Arbor

Citations of this work

No citations found.

Add more citations

References found in this work

On Computable Numbers, with an Application to the Entscheidungsproblem.Alan Turing - 1936 - Proceedings of the London Mathematical Society 42 (1):230-265.
Unprincipled.Gordon Belot - 2024 - Review of Symbolic Logic 17 (2):435-474.
Absolutely No Free Lunches!Gordon Belot - forthcoming - Theoretical Computer Science.
Bayesianism and reliable scientific inquiry.Cory Juhl - 1993 - Philosophy of Science 60 (2):302-319.

View all 6 references / Add more references