Moulin-Frier, Clément and Arsiwalla, Xerxes D. and Puigbò, Jordi-Ysard and Sánchez-Fibla, Martí and Duff, Armin and Verschure, Paul F. M. J.
Mammals bootstrap their cognitive structures through embodied interaction with the world. This raises the question: how the reactive control of initial behavior is recurrently coupled to the high-level symbolic representations they give rise to? We investigate this question in the framework of the "Distributed Adaptive Control" (DAC) cognitive architecture, where we study top-down and bottom-up interactions between low-level reactive control and symbolic rule learning in embodied agents. Reactive behaviors are modeled using a neural allostatic controller, whereas high-level behaviors are modeled using a biologically-grounded memory network reflecting the role of the prefrontal cortex. The interaction of these modules in a closed-loop fashion suggests how symbolic representations might have been shaped from low-level behaviors and recruited for behavior optimization.
@inproceedings{moulinfrier2016topdown,
title = {Top-Down and Bottom-Up Interactions between Low-Level Reactive Control and Symbolic Rule Learning in Embodied Agents},
author = {Moulin-Frier, Cl{\'e}ment and Arsiwalla, Xerxes D. and Puigb{\`o}, Jordi-Ysard and S{\'a}nchez-Fibla, Mart{\'i} and Duff, Armin and Verschure, Paul F. M. J.},
booktitle = {Proceedings of the Workshop on Cognitive Computation: Integrating neural and symbolic approaches. 30th Annual Conference on Neural Information Processing Systems (NIPS 2016)},
year = {2016},
url = {http://clement-moulin-frier.github.io/pdf/moulinfrier2016topdown.pdf}
}
