Fairground2025

On-device training using edge computing allows intelligent devices to learn autonomously while improving latency and security. However, resource constraints remain a challenge. Techniques like weight compression, federated learning, and transfer learning help but may not meet hardware needs. Neuromorphic solutions, such as SNNs, mimic brain-like, energy-efficient, event-driven computation, ideal for robots and implants. They bridge neuroscience and AI, overcoming limitations of traditional ANNs and adapting to dynamic environments. This workshop explores neuromorphic systems and bio-inspired architectures, combined with traditional networks, to optimize computational resources for diverse applications.