Generative Network Intelligence and Optimization Ecosystem – GENIE
GENIE (Generative Network Intelligence and Optimisation Ecosystem) aims to revolutionize network infrastructure management by proposing an approach combining the strengths of Large Language Models (LLMs) and network domain-specific expertise. This novel solution addresses the limitations of conventional techniques, enabling automated, adaptable, and interpretable network management. We aim to translate high-level intentions into network configurations and solutions, ensuring compliance with best practices, performance requirements, and operational constraints. The project will design an LLM pipeline for network management. It will investigate LLMs collaborative and evolutionary strategies to enable parallel processing, including a consensus mechanism, to maintain efficient decision-making and minimize divergent or incorrect outcomes. GENIE's potential impact lies in its ability to deliver significant advances in network infrastructure management while contributing to the development of network automation. This project proposes a unique approach to design innovative, conversational, AI-driven network management.
Project coordination
Nicolas HUIN (Institut de Recherche en Informatique et Systèmes Aléatoires)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
Partnership
IRISA/IMT Institut de Recherche en Informatique et Systèmes Aléatoires
LEAT Laboratoire d'Electronique, Antennes et Télécommunications
IRISA Institut de Recherche en Informatique et Systèmes Aléatoires
EA2118 - L3i LABORATOIRE INFORMATIQUE IMAGE INTERACTION
LabHC Laboratoire Hubert Curien
Help of the ANR 667,834 euros
Beginning and duration of the scientific project:
September 2024
- 48 Months