Demonstrating Volition and Affective Decision Making in Brain-inspired Neural Mode Ls

This paper presents an experiment concerning affective decision making in artificial neural networks (ANNs). The type of neuron used is the virtual G-RAM neuron, which was developed by Aleksander (1990) and implemented on a multimodule neuro-computational software system called Neural Representation Modeller (NRM), created by Aleksander and Dunmall (2000). I have built two neural configurations, each dealing with contrived and worldly scenarios. Each has a set of rules and facts which are pre-learned, and during execution the entity has to decide on an appropriate course of action to take, depending on the situation unfolding in the world. The entity also takes into account its current ‘emotional’ state. I am working on making the decisions largely non-deterministic and random, to create a true freedom of choice. However, as the entity learns about its world, it will develop preferences and habits, and these will introduce a bias in the decision making process. The neural configurations presented in this paper are novel.