Thai readers could see smarter AI via self-organizing infomorphic neurons
A collaboration between the University of Göttingen and the Max Planck Institute for Dynamics and Self-Organization has unveiled “infomorphic neurons” that learn autonomously by mimicking brain-like processes. Published in the Proceedings of the National Academy of Sciences, this work shifts away from traditional supervised neural networks toward self-organizing artificial units. The neurons can determine which inputs matter for learning, reducing the need for constant external guidance.
The human brain operates through decentralized, energy-efficient networks. Biological neurons learn by responding to neighboring cells rather than following rigid, pre-set pathways. Infomorphic neurons imitate this adaptability, selecting learning goals and rules with minimal external control. With self-organization and specialization, these networks promise more robust problem solving in real-world tasks.