Description

Introduction

Agentic systems go beyond reactive behavior by incorporating elements of memory, reasoning, learning, and planning—hallmarks of cognition. To support such advanced capabilities, cognitive architectures provide structured blueprints that model how intelligent agents perceive, decide, and act in complex environments. This course explores the foundational and modern cognitive architectures used to build Agentic AI systems, bridging insights from cognitive science, artificial intelligence, and robotics to design agents that mimic human-like thinking.

Prerequisites

To benefit fully from this course, learners should have:

• A solid understanding of AI and agent-based systems

• Familiarity with concepts such as planning, memory, and goal-oriented behavior

• Basic programming experience (Python preferred)

• Optional: Background in cognitive science or reinforcement learning

Table of Contents

1. Introduction to Cognitive Architectures

• 1.1 What Are Cognitive Architectures?

• 1.2 Role in Agentic AI Development

• 1.3 Cognitive AI vs Symbolic and Subsymbolic AI

2. Core Components of Cognitive Architectures

• 2.1 Perception and Sensory Processing

• 2.2 Working and Long-Term Memory Models

• 2.3 Decision-Making and Action Selection

• 2.4 Learning and Adaptation Mechanisms

3. Classic Cognitive Architectures

• 3.1 Soar: Problem-Space and Rule-Based Reasoning

• 3.2 ACT-R: Modular Memory and Procedural Knowledge

• 3.3 CLARION: Dual-Process Learning and Reasoning

4. Modern Architectures for Agentic Systems

• 4.1 LIDA (Learning Intelligent Distribution Agent)

• 4.2 Sigma Cognitive Framework

• 4.3 ICARUS: Goals, Habits, and Planning

• 4.4 Overview of Hybrid Symbolic-Connectionist Models

5. Architectures in Practice

• 5.1 Cognitive Robotics

• 5.2 Virtual Humans and Game Agents

• 5.3 Human-AI Collaboration and Simulation

6. Integrating LLMs into Cognitive Architectures

• 6.1 Language Models as Cognitive Modules

• 6.2 Embedding Reasoning and Planning via LLMs

• 6.3 Memory-Augmented LLM Agents

7. Evaluation and Design Considerations

• 7.1 Measuring Cognitive Fidelity and Task Performance

• 7.2 Scalability and Real-Time Processing

• 7.3 Modular vs Monolithic Architecture Design

8. Ethical and Philosophical Considerations

• 8.1 Cognitive Modeling vs Real Intelligence

• 8.2 Agency, Autonomy, and Responsibility

• 8.3 Cognitive Transparency and Explainability

Cognitive architectures are key to enabling true agentic behavior in AI systems—supporting memory, learning, reasoning, and adaptive goal pursuit. By modeling the cognitive processes underlying intelligent behavior, these architectures provide a foundation for building more human-aligned and capable agents. This course equips learners to critically evaluate, select, and implement cognitive models in next-gen AI systems across domains such as robotics, education, simulation, and virtual assistance.