A biologically plausible, neural-inspired planning approach which does not solve 'The gourd, the monkey, and the rice' puzzle

This article describes research work aimed at designing a biologically inspired, neural network based computational model for planning. The article extends our previous work (Mastrogiovanni et al. (2010) International Journal of Machine Consciousness, 3, 91–114) by solving problems making the previous approach hard to apply in real-world scenarios. The proposed solution is based on two conjectures on the way the cognitive capabilities of living beings have evolved, and purposely ignores the human-level skills required for solving complex puzzles. In this spirit, we take inspiration from ethology to introduce a new puzzle, henceforth referred to as the 'The gourd, the monkey, and the rice': then, we argue that our planning approach fails to solve the puzzle as well as the real monkeys fail to solve it, and we suggest a biologically plausible explanation for the reason why this happens.