R-IO SUITE: integration of LLM-based AI into a knowledge management and model-driven based platform dedicated to crisis management
Résumé
This article presents how the R-IO SUITE software platform, a decision support system for crisis management entirely based on model-driven engineering principles, considerably benefits from large language model (LLM)-based artificial intelligence (AI). The different components of the R-IO SUITE platform are used to climb the four abstraction layers: data, information, decision and action through interpretation (from data to information), exploitation (from information to decision) and implementation (from decision to action). These transitions between layers are supported by a knowledge base embedding knowledge instances structured according to a crisis management metamodel. From a functional perspective, this platform is fully operational, however, to be able to cover any type of crisis situation, the knowledge base should be enriched, first, from a "resource perspective" (to embed the various available means to deal with any faced situation), and second, from an "issue perspective" (to understand all risks and damage that can appear on a crisis situation). It is not reasonable to consider creating and maintaining such an exhaustive knowledge base. However, the connection of the R-IO SUITE platform with LLM software such as ChatGPT (c) makes it possible, by generating appropriate prompts, to update on-the-fly the knowledge base according to the faced context. This article shows how the LLM AI can provide complementary knowledge to formally fulfil the knowledge base to make it relevant to the faced crisis situation. This article presents the R-IO SUITE as a LLM-empowered model-driven platform to become an extended crisis management supporting system.
Mots clés
Large language model
Artificial intelligence
Model-driven engineering
Decision support system
Knowledge base
Ontologies
Metamodel
Complex event processing
Business process management
Large Language Model Artificial Intelligence Model-Driven Engineering Decision Support System Knowledge base Ontologies Metamodel Complex Event Processing Business Process Management
Large Language Model
Artificial Intelligence
Model-Driven Engineering
Decision Support System
Complex Event Processing
Business Process Management
Domaines
Autre [cs.OH]Origine | Fichiers produits par l'(les) auteur(s) |
---|