Recent Advances in on-Chip Learning with Organic Neuromorphic Circuits

Artificial neural networks power AI applications but require substantial computing resources. Neuromorphic engineering, inspired by the brain, uses organic materials that emulate neuroplasticity, offering low-energy operation and adaptability. Organic electrochemical transistors can be used as neuromorphic devices (ECRAM) to enable efficient edge computing with local, adaptive learning capabilities. We have developed compact neuromorphic circuits that learn through reinforcement and supervised learning, advancing low-power, on-chip training suitable for autonomous systems.