AI Technology Workshop Series: Spiking Neural Networks | AISV.817_W14

---

Welcome to our immersive AI technology workshop series. During these sessions you will be introduced to new and established AI tools that will help you create and manipulate content in new and powerful ways. Each session is led by an industry expert who will guide you through the material and share its real-world implications.

Learning Outcomes

At the conclusion of the workshop, you should be able to

• Describe and discuss the fundamental principles of Spiking Neural Networks (SNNs), including spike-based data representation, key neuron models (e.g., Leaky Integrate-and-Fire), synaptic operations, the current state-of-the-art in neuromorphic hardware, and the realistic short-term and long-term potential of SNNs in the broader AI landscape.
• Explain the differences between Spiking Neural Networks (SNNs) and traditional Artificial Neural Networks (ANNs), articulating the motivational factors for SNN adoption such as energy efficiency and event-driven computation.
• Demonstrate an ability to properly and effectively implement simple SNN applications (e.g., the XOR problem and a more complex use case) using tools like Nengo, understanding how information is encoded and processed through spike trains, and applying different training approaches for SNNs while contrasting them with traditional backpropagation and evaluating the challenges and opportunities in SNN learning.

Topics Include

• Introduction to Spiking Neural Networks: What are SNNs? Visualizing spike-based computation.
• Why SNNs Matter: Motivations (energy efficiency, neuromorphic hardware), real-world applications (robotics, IoT), and a realistic look at when they are most beneficial.
• Core Concepts of SNNs: Spike-based representation (events, timing, frequency), simplified neuron models (LIF), synaptic function, and encoding strategies (rate vs. temporal).
• Hands-On with SNNs (Nengo Demo): Interactive exploration of spike generation, parameter tuning, and visualizing network behavior.
• Problem Solving with SNNs:
  The XOR problem: Understanding challenges with discrete logic.
  Applying SNNs to continuous, real-world analog-like problems.
• Training Spiking Neural Networks: Exploring alternatives to backpropagation (e.g., evolutionary computation, PSO) and understanding the associated challenges.
• The Neuromorphic Landscape: Introduction to key neuromorphic hardware (e.g., Intel Loihi, BrainChip Akida), their architectures, and real-world case studies.
• Current Limitations and Future Outlook: Discussing speed vs. efficiency, challenges in tooling and frameworks, and the short-term and long-term vision for SNN adoption.

Students are required to bring laptops for class exercises