Tableau noir est un modèle informatique largement utilisé dans is a computing model used in intelligence artificielle where multiple agents or processes collaborate to solve complex problems by sharing a common base de connaissances, referred to as the ‘blackboard.’
In this architecture, the blackboard acts as a shared workspace where different knowledge sources can contribute information, hypotheses, or solutions. Each agent or module, known as a ‘knowledge source,’ has its own expertise and can read from and write to the blackboard. This allows for a flexible and dynamic problem-solving approach, as different agents can contribute at different stages of the process.
The blackboard architecture is particularly effective in scenarios where problems are too complex for a single agent to solve independently. For example, in traitement du langage naturel, various agents might handle syntax, semantics, and discourse analysis, contributing their findings to the blackboard to create a coherent understanding of the text.
Composants typiques d'une architecture à tableau noir :
- Sources de connaissances : Le référentiel central où sont stockés les données, hypothèses et solutions.
- Composant de contrôle : Independent agents or modules that access the blackboard, contribute knowledge, and make inferences.
- Composant de contrôle : Manages the activity of knowledge sources, determining which source can act and when, often based on the current state du tableau noir.
Cette architecture est polyvalente et a été appliquée dans divers domaines, notamment reconnaissance vocale, image processing, and expert systems, showcasing its ability to handle complex, multifaceted problems through collaboration.