Neuromorphic Hardware – A System Perspective

Overview

Explore neuromorphic computing principles, systems like SpiNNaker2, and future developments in energy-efficient AI hardware with Johannes Partzsch from TU Dresden.

Syllabus

• Introduction to Neuromorphic Computing
Overview of Neuromorphic Computing
Historical Development and Inspiration from Biology
Key Differences from Conventional Computing Architectures
• Fundamental Principles of Neuromorphic Systems
Spiking Neural Networks (SNNs)
Neurons and Synapses: Biological Inspiration
Event-Driven Processing
• Neuromorphic Hardware Architectures
Key Hardware Components and Their Functions
Comparison of Neuromorphic and Traditional Hardware
Challenges in Neuromorphic System Design
• In-Depth Case Study: SpiNNaker2
Architecture and Design Principles
Scalability and Modular Design
Programming and Software Ecosystem
• Energy Efficiency in AI Hardware
Importance of Energy Efficiency in AI
Techniques for Reducing Power Consumption in Neuromorphic Systems
Comparison with Conventional AI Hardware
• Applications of Neuromorphic Computing
Real-World Use Cases
Advantages and Limitations
Future Potential and Emerging Trends
• Future Developments in Neuromorphic Systems
Trends in Technology and Fabrication
Prospective Innovations and Research Directions
Integration with Other Computing Paradigms (e.g., Quantum Computing)
• Course Summary and Implications for AI Hardware
Recap of Key Concepts
Discussion: Impact of Neuromorphic Hardware on AI Development
Ethical and Societal Considerations
• Final Project/Assessment
Design and Evaluation of a Simple Neuromorphic System
Analysis of Energy-Efficiency Trade-offs in Neuromorphic vs. Traditional AI Hardware

Subjects

Computer Science