« Audit algorithmique » : différence entre les versions
(Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Algorithmic auditing''' '''shadow auditing''' Also known as shadow... ») |
m (Remplacement de texte — « [http » par « * [http ») |
||
Ligne 15 : | Ligne 15 : | ||
<small> | <small> | ||
[https://www.forbes.com/sites/insights-intelai/2019/05/22/a-glossary-for-next-generation-ai/?sh=40a218e477d9 Source : forbes ] | * [https://www.forbes.com/sites/insights-intelai/2019/05/22/a-glossary-for-next-generation-ai/?sh=40a218e477d9 Source : forbes ] | ||
[[Catégorie:vocabulary]] | [[Catégorie:vocabulary]] |
Version du 4 février 2021 à 23:25
en construction
Définition
XXXXXXXXX
Français
XXXXXXXXX
Anglais
Algorithmic auditing
shadow auditing
Also known as shadow auditing. The process is used to identify AI “blind spots.” As concerns grow about the hidden bias in AI systems, algorithmic auditing is a way to seek out flaws in structural design, coding and training data sets and to assess the system for consistency, transparency, accuracy and fairness. Such auditing is being used to detect bias in AI tools used to make decisions in financial services, the criminal justice system and hiring.
Contributeurs: Imane Meziani, wiki