The no-free-lunch theorems of supervised learning.
The no-free-lunch theorems promote a skeptical conclusion that all possible machine learning algorithms equally lack justification. But how could this leave room for a learning theory, that shows that some algorithms are better than others? Drawing parallels to the philosophy of induction, we point...
| Publicado en: | Synthese Vol. 199; no. 3/4; pp. 9979 - 10016 |
|---|---|
| Autores principales: | , |
| Formato: | Artículo |
| Publicado: |
Springer Nature
Dec2021
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=hlh&AN=154096882&site=ehost-live header: @attributes: shortDbName: hlh uiTerm: 154096882 longDbName: Humanities International Complete uiTag: AN controlInfo: bkinfo: jinfo: jid: 00397857 4LI jtl: Synthese issn: 00397857 maglogo: N pubinfo: dt: Dec2021 vid: 199 iid: 3/4 pid: 237 pub: Springer Nature artinfo: ui: 154096882 10.1007/s11229-021-03233-1 ppf: 9979 ppct: 37 formats: fmt: – @attributes: type: T – @attributes: type: P size: 637KB tig: atl: The no-free-lunch theorems of supervised learning. aug: au: Sterkenburg, Tom F. Grünwald, Peter D. affil: Munich Center for Mathematical Philosophy, LMU Munich, Munich, Germany Machine Learning Group, Centrum Wiskunde & Informatica, Amsterdam, The Netherlands Mathematical Institute, Leiden University, Leiden, The Netherlands su: Machine learning Algorithms sug: subj: Machine learning Algorithms keyword: No-free-lunch theorems Problem of induction ab: The no-free-lunch theorems promote a skeptical conclusion that all possible machine learning algorithms equally lack justification. But how could this leave room for a learning theory, that shows that some algorithms are better than others? Drawing parallels to the philosophy of induction, we point out that the no-free-lunch results presuppose a conception of learning algorithms as purely data-driven. On this conception, every algorithm must have an inherent inductive bias, that wants justification. We argue that many standard learning algorithms should rather be understood as model-dependent: in each application they also require for input a model, representing a bias. Generic algorithms themselves, they can be given a model-relative justification. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: Y custom: Synthese is a copyright of Springer, 2021. All Rights Reserved. item: Synthese holder: Springer Nature dt: @attributes: year: 2021 holdings: @attributes: islocal: N |
|---|