Collective brain surrogates

Within the framework of ensemble methods, we investigate on a compatible learning scheme, denoted as learning by gossip, with the aim of assessing its feasibility when facing a rather complex target function. Compatibility is expressed in terms of probability that the learned function could be actually at the basis of the observed training set, hence an explanation of it. Feasibility is in terms of the related Mean Square Error (MSE) on test sets. Elaborating on ways to improve the performance of the learning scheme, we assessed its reliability, efficacy and exploitability, via numerical tools that play the role of learning, educating, feeling and achieving consciousness in a virtual society. We base or conclusions on both theoretical and numerical arguments that are tossed on a well known benchmark. We devote a large space to provide graphical evidence to some conclusions that may be exploited in the field of both connected societies and cooperative computation frameworks.