Journal article Open Access

Learning programs is better than learning dynamics: A programmable neural network hierarchical architecture in a multi-task scenario

Francesco Donnarumma; Roberto Prevete; Andrea de Giorgio; Guglielmo Montone; Giovanni Pezzulo

Distributed and hierarchical models of control are nowadays popular in computational modeling and robotics. In the artificial neural network literature, complex behaviors can be produced by composing elementary building blocks or motor primitives, possibly organized in a layered structure. However, it is still unknown how the brain learns and encodes multiple motor primitives, and how it rapidly reassembles, sequences and switches them by exerting cognitive control. In this paper we advance a novel proposal, a hierarchical programmable neural network architecture, based on the notion of programmability and an interpreter-programmer computational scheme. In this approach, complex (and novel) behaviors can be acquired by embedding multiple modules (motor primitives) in a single, multi-purpose neural network. This is supported by recent theories of brain functioning in which skilled behaviors can be generated by combining functional different primitives embedded in “reusable” areas of “recycled” neurons. Such neuronal substrate supports flexible cognitive control, too. Modules are seen as interpreters of behaviors having controlling input parameters, or programs that encode structures of networks to be interpreted. Flexible cognitive control can be exerted by a programmer module feeding the interpreters with appropriate input parameters, without modifying connectivity. Our results in a multiple T -maze robotic scenario show how this computational framework provides a robust, scalable and flexible scheme that can be iterated at different hierarchical layers permitting to learn, encode and control multiple qualitatively different behaviors.

Files (5.5 MB)
Name Size
Learning programs is better than learning.pdf
md5:8b065c35f9ddafa60a16464441b67922
5.5 MB Download
45
40
views
downloads
Views 45
Downloads 40
Data volume 218.1 MB
Unique views 41
Unique downloads 38

Share

Cite as