ABSTRACT
The ability to assess the similarity of objects in the world is fundamentally important to our survival. A variety of theories have been proposed for modeling human similarity judgments. Most of these theories involve comparing the sets of features of different objects to determine the overlap between them. But most of the theories also ignore the structure of the objects and relationships between the parts. Goldstone’s SIAM system uses a connectionist architecture to create correspondences between objects and their features in different scenes (Goldstone, 1994). Excitatory connections reinforce coherent mappings between objects (e.g. ObjectA to ObjectC and ObjectB to ObjectD). Inhibitory connections fight against redundant or contradictory mappings (e.g. ObjectA to ObjectC and Object A to ObjectD). Likewise, connections between the features of objects either support or inhibit each other and the corresponding object-object connections. Siam’s connectionist architecture allows it to take into account the structure of the scenes and the objects as well as the similarity of the features.
