Aislinn Cognitive Architecture

IMPORTANT DISCLAIMER: This project is currently under active development and is not yet complete. Much of the code is being ported from V1 and reorganized into a new structure. Most components have not been tested. Expect frequent changes and potential instability as development continues.

AISLINN: Artificial Intelligence Systems Leveraging Intuition and Neurocognitive Nuance

Aislinn is a modular, extensible cognitive architecture designed for creating intelligent agents with human-like cognitive capabilities. Unlike many academic cognitive architectures focused on modeling specific aspects of cognition for research, Aislinn is built for practical applications, providing a comprehensive framework for general-purpose AI agents that can reason, learn, plan, and interact naturally with humans and environments.

Explore the API Documentation

Key Features

• Chunk-Based Memory System: Inspired by cognitive science but engineered for practical applications
• Spreading Activation Networks: Dynamic activation flows through knowledge structures based on relevance
• Emotional Modeling: Implements the PAD (Pleasure-Arousal-Dominance) emotional model for realistic affective states
• Goal Management System: Hierarchical goal structures with sophisticated priority management
• Procedural Knowledge: Flexible representation of actions and procedures for accomplishing goals
• Context-Aware Processing: Maintains situational awareness that influences cognitive processes
• Modular Design: Extensible components that can be enhanced or replaced independently

Core Components

Memory Systems

• Working Memory: Human-like capacity limitations with interference and decay
• Long-term Memory: Persistent storage with activation-based retrieval
• Associative Memory: Relationship-based connections between knowledge chunks
• Procedural Memory: Action sequences with preconditions and effects

Goal Management

• Goal Templates: Reusable patterns for goal instantiation
• Goal Hierarchy: Complex goals decompose into subgoals
• Goal Selection: Context-sensitive priority calculation
• Goal Monitoring: Tracks progress and completion criteria

Procedural Execution

• Procedure Matching: Identifies procedures that can achieve specific goals
• Execution Engine: Manages the running of procedures with monitoring
• Error Handling: Sophisticated retry and recovery strategies
• Resource Management: Tracks and allocates resources needed for procedures

Cognitive Context

• Environmental Context: Awareness of physical surroundings
• Internal Context: Mental and emotional states
• Social Context: Understanding of social dynamics and relationships
• Task Context: Current activities and their progress
• Emotional Context: PAD-based emotion modeling affecting decision-making and behavior

Architecture Design

Aislinn is built as a modular system where components communicate through well-defined interfaces. This allows:

• Extensibility: Add new cognitive capabilities without disrupting existing ones
• Customization: Tailor the architecture for specific application domains
• Experimentation: Test different approaches to specific cognitive functions
• Scalability: Deploy configurations ranging from lightweight to comprehensive

Getting Started

Prerequisites

• .NET 9
• Any IDE supporting C# (Visual Studio, VS Code with C# extensions, etc.)

Installation

1. Clone the repository:

   git clone https://github.com/jmlothian/AislinnV2.git
   

2. Open the solution in your preferred IDE

3. Build the solution:

   dotnet build
   

Basic Usage

// Initialize core services
var chunkStore = new ChunkStore();
var associationStore = new AssociationStore();
var timeManager = new CognitiveTimeManager();
var activationModel = new ActRActivationModel(timeManager);

// Create memory system
var memorySystem = new CognitiveMemorySystem(
    chunkStore,
    associationStore,
    activationModel,
    timeManager);

// Initialize goal management
var goalManager = new GoalManagementService(
    chunkStore,
    associationStore,
    new ChunkActivationService(chunkStore, associationStore, activationModel));

// Create a new goal
var goal = await goalManager.CreateGoalTemplateAsync(
    "LearnTopic",
    0.8,
    new List<string> { "topic" });

// Instantiate the goal
var learningGoal = await goalManager.InstantiateGoalAsync(
    goal.ID,
    new Dictionary<string, object> { { "topic", "AI Architecture" } });

// Start cognitive cycle
await memorySystem.StartWorkingMemoryRefresh();

Applications

Aislinn is designed for creating intelligent agents across various domains:

• Virtual Assistants: Sophisticated agents with common-sense reasoning
• Game Characters: NPCs with believable behavior and memory
• Simulation Agents: Entities that behave realistically in virtual environments
• Robotic Control: Cognitive layer for robot decision-making
• Interactive Storytelling: Characters that adapt to narrative context

Development Status

This project is a port and reorganization of Aislinn V1. The current codebase represents the architecture and structure of the system, but many components are still being integrated and tested. Contributors should note:

• Many classes have been ported but not yet tested in the new organization
• Interfaces may change as integration continues

Acknowledgments

While Aislinn draws inspiration from cognitive architectures like ACT-R, Soar, and CLARION, it prioritizes practical application over strict cognitive modeling, providing a flexible foundation for building next-generation intelligent agents.