Neural Networks (IJCNN), The 2012 International Joint Conference on

DOI: 10.1109/IJCNN.2012.6252492

Abstract—As computers approach the physical limits of in- formation storable in memory, new methods will be needed to further improve information storage and retrieval. We propose a quantum inspired vector based approach, which offers a contextually dependent mapping from the subsymbolic to the symbolic representations of information. If implemented computationally, this approach would provide exceptionally high density of information storage, without the traditionally required physical increase in storage capacity. The approach is inspired by the structure of human memory and incorporates elements of Gaerdenfors’ Conceptual Space approach and Humphreys et al.’s matrix model of memory.

The key components I find interesting are the relationship between symbolic and subsymbolic levels, and its movement towards a distributed overlaid model of memory, that causes excitations in related symbolic entities/terms.  The matrix notation allows for a representation of context as a matrix of features, which via tensor product formalism,  allows for memory to be a distributed process, with activations of related terms.

This is of interest to me, as we can start to utilise visual features in a computational model of subsymbolic components contributing to memory recall via similar inputs of features.

While this is strongly related to an NN associated matrix approach, it makes the relationships between the components explicit, and is thus a candidate as a cognitive model of priming in expert elicitation sessions.  This approach can then be used to modulate user interfaces in virtual world elicitation systems.