Towards Turning MLOps into a Continuous Learning Process - IMT Mines Alès Accéder directement au contenu
Poster De Conférence Année : 2024

Towards Turning MLOps into a Continuous Learning Process

Charbel Daoud
  • Fonction : Auteur
  • PersonId : 1393766
Julie Boiché
  • Fonction : Auteur
  • PersonId : 1393767
Christelle Urtado
Sylvain Vauttier

Résumé

MLOps has emerged in the past decade, aiming to define and implement best practices for developing and deploying machine learning (ML) models. Researchers have since been exploring and applying MLOps to various use cases, thus contributing to a better understanding and definition of its requirements and practical implementation. This poster examines the challenges faced by MLOps research and investigates the advancements and challenges addressed by ML paradigms in general. By merging these two areas of work, our preliminary idea is to introduce Continuous Learning as a new pillar for MLOps. Indeed, Continuous Learning will enhance MLOps with the capability to learn from past choices, reduce the need for constant retraining of models, and mitigate issues related to both data and concept drifts.
Fichier principal
Vignette du fichier
GDR_GPL_2024_Demos__Posters_paper_7.pdf (994.86 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04617819 , version 1 (19-06-2024)

Identifiants

  • HAL Id : hal-04617819 , version 1

Citer

Charbel Daoud, Danielle Azar, Julie Boiché, Christelle Urtado, Sylvain Vauttier. Towards Turning MLOps into a Continuous Learning Process. Journées nationales du Groupement de Recherche Génie de la Programmation et du Logiciel (GdR GPL), Jun 2024, Strasbourg, France. 2024. ⟨hal-04617819⟩
0 Consultations
0 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More