Functional concept proxies and the actually smart Hans problem: what’s special about deep neural networks in science

Simple item page
dc.contributor.authorBoge, Florian J.
dc.date.accessioned2025-02-24T13:41:43Z
dc.date.available2025-02-24T13:41:43Z
dc.date.issued2024-12-27
dc.description.abstractDeep Neural Networks (DNNs) are becoming increasingly important as scientific tools, as they excel in various scientific applications beyond what was considered possible. Yet from a certain vantage point, they are nothing but parametrized functions of some data vector , and their ‘learning’ is nothing but an iterative, algorithmic fitting of the parameters to data. Hence, what could be special about them as a scientific tool or model? I will here suggest an integrated perspective that mediates between extremes, by arguing that what makes DNNs in science special is their ability to develop functional concept proxies (FCPs): Substructures that occasionally provide them with abilities that correspond to those facilitated by concepts in human reasoning. Furthermore, I will argue that this introduces a problem that has so far barely been recognized by practitioners and philosophers alike: That DNNs may succeed on some vast and unwieldy data sets because they develop FCPs for features that are not transparent to human researchers. The resulting breach between scientific success and human understanding I call the ‘Actually Smart Hans Problem’.en
dc.identifier.urihttp://hdl.handle.net/2003/43496
dc.identifier.urihttp://dx.doi.org/10.17877/DE290R-25329
dc.language.isoen
dc.relation.ispartofseriesSynthese; 203(1)
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectDeep Neural Networksen
dc.subjectConceptsen
dc.subjectReasoningen
dc.subjectClever Hans Problemen
dc.subjectAutomated scienceen
dc.subject.ddc100
dc.subject.rswkTiefes neuronales Netz
dc.subject.rswkDeep Neural Networkde
dc.subject.rswkKonzeptionde
dc.subject.rswkBeweisführungde
dc.subject.rswkClever Hans Phenomenonde
dc.subject.rswkAutomatisierungstechnikde
dc.titleFunctional concept proxies and the actually smart Hans problem: what’s special about deep neural networks in scienceen
dc.typeText
dc.type.publicationtypeResearchArticle
dcterms.accessRightsopen access
eldorado.secondarypublicationtrue
eldorado.secondarypublication.primarycitationBoge, F.J. (2024) ‘Functional concept proxies and the actually smart Hans problem: what’s special about deep neural networks in science’, Synthese, 203(1). Available at: https://doi.org/10.1007/s11229-023-04440-8
eldorado.secondarypublication.primaryidentifierhttps://doi.org/10.1007/s11229-023-04440-8

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
s11229-023-04440-8.pdf
Size:
2.07 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
4.82 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Institut für Philosophie