Biologically inspired neural networks for the control of embodied agents
Razvan V. Florian
This paper reviews models of neural networks suitable for the control of artificial intelligent agents interacting continuously with an environment. We first define the characteristics needed by those neural networks. We review several classes of neural models and compare them in respect of their suitability for embodied agent control. Among the classes of neural network models amenable to large scale computer simulation, recurrent spiking neural networks result to be the most suited for this task. We present next several models of spiking neural networks and review models of unsupervised learning for them. Finally, we review current implementations of spiking neural networks as controllers for embodied agents.