Challenging the resource-intensive "bigger is better" trend of Large Language Models (LLMs), this webinar introduced the Expressive Neural Network (ENN). This novel architecture rethinks

Challenging the resource-intensive "bigger is better" trend of Large Language Models (LLMs), this webinar introduced the Expressive Neural Network (ENN). This novel architecture rethinks activation functions using classical signal processing (Discrete Cosine Transform), resulting in enhanced flexibility, faster convergence, and significantly smaller, more energy-efficient AI models.

ENNs demonstrate how traditional signal processing can inspire next-generation AI, enabling efficient edge computing in resource-constrained environments. This approach prioritizes expressiveness over sheer size for future neural networks. The webinar featured speaker Ana Pérez-Neira, moderated by Ian F. Akyildiz and Alessia Magliarditi. It was organized by the ITU Journal on Future and Evolving Technologies (ITU-J FET).