The Little-Hopfield neural network programmed with Horn clauses is studied. We argue that the energy landscape of the system, corresponding to the inconsistency function for logical interpretations of the sets of Horn clauses, has minimal ruggedness. This is supported by computer simulations.
@article{arxiv.0805.0197,
title = {Flatness of the Energy Landscape for Horn Clauses},
author = {Saratha Sathasivam and Wan Ahmad Tajuddin Wan Abdullah},
journal= {arXiv preprint arXiv:0805.0197},
year = {2008}
}