Recruitment learning

This monograph sets out recruitment learning, a theoretical model of rapid learning in the neocortex. Rather than relying on attractor networks, the usual account of both long- and short-term memory, the model explains how a single exposure can bind together any set of features that appear at the same time, a process described as one-shot learning or chunking.

The opening chapters trace the intellectual history behind the idea. They give particular attention to the contributions of D.O. Hebb, W.A. Wickelgren, J.A. Feldman, L.G. Valiant and L. Shastri, showing how their work led to this line of thinking.

Next, the author works through a detailed mathematical analysis of the model, arguing that recruitment is a credible mechanism for forming memories in the neocortex. A later section carries these ideas over to modern spiking neuron models and proposes dynamic synchronization as a possible answer to the binding problem. The book also considers what role adult neurogenesis might play in recruitment.