The effect of training methodology on knowledge representation in categorization

Abstract

Category representations can be broadly classified as containing within–category information or between–category information. Although such representational differences can have a profound impact on decision–making, relatively little is known about the factors contributing to the development and generalizability of different types of category representations. These issues are addressed by investigating the impact of training methodology and category structures using a traditional empirical approach as well as the novel adaptation of computational modeling techniques from the machine learning literature. Experiment 1 focused on rule–based (RB) category structures thought to promote between–category representations. Participants learned two sets of two categories during training and were subsequently tested on a novel categorization problem using the training categories. Classification training resulted in a bias toward between–category representations whereas concept training resulted in a bias toward within–category representations. Experiment 2 focused on information-integration (II) category structures thought to promote within–category representations. With II structures, there was a bias toward within–category representations regardless of training methodology. Furthermore, in both experiments, computational modeling suggests that only within–category representations could support generalization during the test phase. These data suggest that within–category representations may be dominant and more robust for supporting the reconfiguration of current knowledge to support generalization.

The impact of category structure and training methodology on learning and generalizing within-category representations

“…In contrast, II structures are those in which information from multiple dimensions needs to be integrated prior to making a categorization response. Unlike RB structures, II structures generally promote within-category representations (Ashby & Waldron, 1999;Hélie et al, 2017;Thomas, 1998). Again, even when classification is the goal, RB structures would be expected to promote between-category representations, whereas II structures would be expected to promote withincategory representations.…”

“…Both classification and concept training are active tasks and have the goal of classification on a trial-by-trial basis. Concept training, however, has been argued to shift the emphasis from between-category differences to within-category similarities (Casale & Ashby, 2008;Hélie, Shamloo, & Ell, 2017).…”

“…RB structures can be learned using logical rules. Although logical rules can be based on either within-or between-category representations (e.g., large or larger than), the subset of logical rules learned with RB structures tends to depend upon between-category representations (Casale, Roeder, & Ashby, 2012;Ell, Ing, & Maddox, 2009;Hélie et al, 2017).…”

“…Indeed, within-category representations can support generalization to novel tasks (Chin-Parker & Ross, 2002). Within-category representations are also better able than between-category representations to support the reconfiguration of categorical knowledge (Hélie et al, 2017;Hoffman & Rehder, 2010).…”

“…Although participants often demonstrate an initial bias toward between-category representations (Ashby, Queller, & Berretty, 1999;Ell & Ashby, 2006;Medin, Wattenmaker, & Hampson, 1987;Smith, Beran, Crossley, Boomer, & Ashby, 2010), within-category representations may be a common outcome of interacting with categories (e.g., Anderson & Fincham, 1996;Hélie et al, 2017;Hoffman & Rehder, 2010;Thomas, 1998). Previous work suggests that numerous methodological factors can promote within-category representations, but there is variability in how within-category representations were measured, if measured at all.…”