Skip to main content

Architecture

One of the main objectives of the VALAWAI project is to build upon the cognitive model of the Global Neuronal Workspace theory (GNW) to develop a computational cognitive architecture for modeling value awareness in AI. This architecture provides a foundation for creating value-aware applications and promoting responsible AI development.

This section provides a comprehensive overview of the VALAWAI architecture, covering the following key areas:

  • Neural Model: This section details the underlying neural models that inspire the VALAWAI architecture. It describes the Global Neuronal Workspace (GNW) model and the Reflective Global Neuronal Workspace (RGNW) model, explaining how these models inform the project's approach to value awareness. It also discusses the key principles and mechanisms of these models relevant to the VALAWAI implementation.

  • Value Awareness in VALAWAI: This section clearly defines what value awareness means within the specific context of the VALAWAI project. It explores the project's interpretation of values, how they are represented, and how they influence decision-making within the AI system.
    It also clarifies the scope and limitations of the project's approach to value awareness.

  • VALAWAI Architecture: This section provides a detailed description of the VALAWAI architecture itself.
    It outlines the different components of the architecture, their interactions, and how they contribute to value-aware behavior. Diagrams and visual representations are included to illustrate the architecture's structure and data flow. This section also explains how the architecture incorporates the principles of the chosen neural models (GNW/RGNW).

  • Implementations: This section showcases the different implementations of the VALAWAI architecture.
    It provides concrete examples of how the architecture has been applied in various contexts, demonstrating its capabilities and flexibility. Code snippets, usage examples, and links to relevant repositories are included where appropriate. This section also highlights best practices for implementing and extending the architecture.