An open commentary just appeared on Clarks BBS article “Whatever next? Predictive brains, situated agents, and the future of cognitive science”. In brief, the commentary discusses the more general mechanism of internal simulation:
… But there may be more to [internal models] then only prediction. First of all, engaged (and maybe grounded as sensorimotor circuits) internal models may serve an inverse function in motor control, i.e., coming up with motor control commands when given a certain goal. Secondly, as animals and humans have a wide variety of redundant sensors, such internal models should exploit the redundancy and integrate the noisy contributions of multiple sensors.
As we know today, such internal models are not only serving one single function, but internal models are recruited in service for diverse function (Anderson, 2010), e.g., perception, to understand the actions performed by somebody else, or in planning ahead. The internal models are recruited in an internal simulation (Hesslow, 2002) – and central is their predictive function: in planning ahead they are used to simulate possible consequences of actions and then to choose only a suitable one.
The complete commentary can be found here.
A new article appeared in the journal Frontiers in Cognition. It is about how internal models are recruited in prediction and demonstrates a simple neural network that can be used in prediction as well as motor control.
Even comparatively simple, reactive systems are able to control complex motor tasks, such as hexapod walking on unpredictable substrate. The capability of such a controller can be improved by introducing internal models of the body and of parts of the environment. Such internal models can be applied as inverse models, as forward models or to solve the problem of sensor fusion. Usually, separate models are used for these functions. Furthermore, separate models are used to solve different tasks. Here we concentrate on internal models of the body as the brain considers its own body the most important part of the world. The model proposed is formed by a recurrent neural network with the property of pattern completion. The model shows a hierarchical structure but nonetheless comprises a holistic system. One and the same model can be used as a forward model, as an inverse model, for sensor fusion, and, with a simple expansion, as a model to internally simulate (new) behaviors to be used for prediction. The model embraces the geometrical constraints of a complex body with many redundant degrees of freedom, and allows finding geometrically possible solutions. To control behavior such as walking, climbing or reaching, this body model is complemented by a number of simple reactive procedures together forming a procedural memory. In this article, we illustrate the functioning of this network. To this end we present examples for solutions of the forward function and the inverse function, and explain how the complete network might be used for predictive purposes. The model is assumed to be “innate,” so learning the parameters of the model is not (yet) considered.
The paper appeared as open access and can be reached following this link.