Watch a technical conference presentation from tinyML Asia 2023 exploring modular event-driven spiking neural architectures for low-power neuromorphic systems. Discover how the Optimized Deep Event-driven Spiking neural network Architecture (ODESA) enables simultaneous learning of hierarchical spatio-temporal features at multiple time scales through local synaptic and threshold adaptation rules. Learn about ODESA's unique online learning capabilities that operate without error back-propagation or gradient calculations, instead utilizing simple local adaptive selection thresholds for efficient neuronal resource allocation. Examine the efficient FPGA hardware implementation and understand how the binary event-based communication between layers enables asynchronous, low-power always-on learning systems, providing valuable insights for future neuromorphic architecture development.