Modeling Human Reading in Conceptual Networks for Text Representation and Comparison

Although machines perform much better than human beings in most of the tasks, it is not the case of natural language processing. Computational linguistics systems use to rely on mathematical and statistical formalisms, which are efficient and useful but far from human procedures and therefore not so skilled. This paper proposes a computational model of natural language reading, called cognitive reading indexing model (CRIM), inspired by some aspects of human cognition, trying to become as more psychologically plausible as possible. The model relies on a semantic neural network and it does not produce vectors but nets of activated concepts as text representations. Based on these representations, an efficient measure of semantic similarity is also defined. The system is not only suitable to human reading modeling but also it can be used in natural language processing applications since results point out that the system improves the performance of other traditional language representations.