Transparence des algorithmes face à l'Open Data : Quel statut pour les données d'apprentissage ?

Abstract : Public administration's reliance on software algorithms brings the risk of obfuscating the factual and legal basis of administrative decisions. In accordance with the principles of open data, new rules have been introduced to preserve the transparency of algorithmic decisions; yet nothing has been said concerning the data that will influence these algorithms. This article investigates the issues stemming from the use of algorithms in the public administration, and underlines the need to guarantee not only access to the source code of these algorithms, but also access to the databases on which they have been trained.
Document type :
Journal articles
Complete list of metadatas

Cited literature [11 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02046703
Contributor : Primavera de Filippi <>
Submitted on : Friday, February 22, 2019 - 6:37:29 PM
Last modification on : Wednesday, February 27, 2019 - 1:19:11 AM
Long-term archiving on : Thursday, May 23, 2019 - 3:34:52 PM

File

61 - (2018) - RFAP - Transpare...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02046703, version 1

Collections

Citation

Danièle Bourcier, Primavera de Filippi. Transparence des algorithmes face à l'Open Data : Quel statut pour les données d'apprentissage ?. Revue française d'administration publique, ENA, 2018. ⟨hal-02046703⟩

Share

Metrics

Record views

21

Files downloads

245