Neuromolecular computing: mechanisms and architectures

The contrasting capabilities of biological systems and digital computers suggest radically different modes of information processing. The difference is connected to a fundamental tradeoff among programmability, computational efficiency, and evolutionary adaptability. Biological systems operate on the efficiency-adaptability side of this tradeoff. Biomolecular geometry (computing by shape) replaces electronic logic in this domain. The author describes a brain model that accommodates a hierarchy of processes built on top of shape-based molecular computing, including electrochemical processes in single neurons and collective processes in neural networks. The model serves as a computer architecture capable of recruiting molecular mechanisms for novel forms of computing.<>