Figure 5 | Scientific Reports

Figure 5

From: Neural network based formation of cognitive maps of semantic spaces and the putative emergence of abstract concepts

Figure 5The alternative text for this image may have been generated using AI.

Different scalings of cognitive maps. Shown are MDS projections of SR matrices with different discount factors \(\gamma\). (A) For a low discount factor \(\gamma =0.3\), the resulting map is most fine-grained and detailed with little clustering (GDV: \(-0.322\)). (B) A medium discount factor of \(\gamma =0.7\) results in an intermediate scale with more clustering (GDV: \(-0.355\) ) compared to (A). (C) The largest discount factor \(\gamma =1.0\) results in the most coarse-grained map. Here, individual animal species are no longer distinguishable, but instead form separated, dense clusters possibly enabling the emergence of more abstract concepts in subsequent processing stages, i.e. the taxonomic animal classes mammals (blue, purple), insects (orange), and amphibians (yellow) (GDV: \(-0.403\)). Note that, a GDV of \(-1.0\) indicates perfect clustering, whereas a GDV of 0.0 indicates no clustering at all. Note that, the axes in the MDS plots are in arbitrary units and have no particular meaning other than illustrating the relative positions, i.e. similarities, of all objects.

Back to article page