Sparse Self-Federated Learning for Energy-Efficient Cooperative Intelligence in Society 5.0
Date:
I presented our paper “Sparse Self-Federated Learning for Energy-Efficient Cooperative Intelligence in Society 5.0” at IJCNN 2025.
The talk introduced a resource-aware cooperative learning approach that combines self-organized federations with neural-network sparsification. The objective is to reduce communication and computational costs while retaining effective collaborative learning across heterogeneous and resource-constrained devices.
The work was presented in the context of green and sustainable AI, with particular attention to the scalability of federated learning in large IoT and edge-computing ecosystems.