SUSTAIN: a network model of category learning.

By Bradley C. Love, Douglas L. Medin, Todd M. Gureckis. Psychology Review, 2004

801 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/eed8bc21cec42d1adb45a1a8f00db7b60aed6137

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Similarity-Dissimilarity Competition in Disjunctive Classification Tasks (2013, background): “SUSTAIN occupies a position between exemplar models that encode every observed example from all presented categories and prototype models that summarize a category by a single central tendency.”
Multi-Prototype Vector-Space Models of Word Meaning (2010, methodology): “Traditionally, word types are represented by a single vector of contextual features derived from cooccurrence information, and semantic similarity is computed using some measure of vector distance (Lee, 1999; Lowe, 2001).”
Formal Approaches in Categorization: Prototype models of categorization: basic formulation, predictions, and limitations (2011, background): “SUSTAIN has been successfully applied to a broad range of data, suggesting that prototypes may provide the best solution to a classification problem, even when an exemplar-based solution is possible.”
A Rational Analysis of Rule-Based Concept Learning (2008, methodology): “Our DNF representations are in some ways more similar to the cluster-based representations of the SUSTAIN model (Love et al., 2004) than they are to representations in RULEX; conjunctive blocks of RRDNF formulae are analogous to the clusters that SUSTAIN learns, with features that are ommitted from ”
Comparing continual task learning in minds and machines (2018, background): “…processes allow decision-relevant features to be appropriately clustered into categories, explaining why feedback-driven category learning is easier when stimuli already differ on fewer dimensions (as in “ rule-based ” tasks), or exhibit decision-relevant features that are a priori salient (36).”
Cognitive control over learning: creating, clustering, and generalizing task-set structure. (2013, positive): “Other models rely on different systems (such as exemplar, within category clusters, and attentional learning) to allow learning of rules less dependent on similarity (Kruschke, 2011; Love et al., 2004; Hahn et al., 2010).”
Evaluating Amazon's Mechanical Turk as a Tool for Experimental Behavioral Research (2013, methodology): “In addition, the quantitative (rather than qualitative) shape of the learning curves has been used to test and differentiate models [43,44].”
Concepts and categories: Memory, meaning, and metaphysics (2012, background): “Other work suggests that having a cohesive category structure is more important for inference learning than it is for classification (Yamauchi & Markman, 1998, 2000a, 2000b; Yamauchi, Love, & Markman, 2002; for modeling implications see Love, Markman, & Yamauchi, 2000; Love et al., 2004).”
Generalization and similarity in exemplar models of categorization: Insights from machine learning (2008, background): “The similarity kernel forms the basis of many categorization models (Kruschke, 1992; Love et al., 2004; Nosofsky, 1986).”
Towards more human-like concept learning in machines : compositionality, causality, and learning-to-learn (2014, methodology): “, Schyns, 1991; Rehder & Murphy, 2003), the approach has been used to implement forms of prototype models (Gluck & Bower, 1988), exemplar models (Kruschke, 1992), and clustering models (Love et al., 2004).”
Deep Neural Networks Predict Category Typicality Ratings for Images (2015, background): “These advances should interest the cognitive science community, especially since categorization is a foundational problem and some leading models are neural networks (e.g., Kruschke, 1992; Love, Medin, & Gureckis, 2004).”
A probabilistic account of exemplar and category generation. (2013, background): “For example, the GCM (Nosofsky, 1986), ALCOVE (Kruschke, 1992), and SUSTAIN (Love et al., 2004) are three wellknown examples of discriminative models and there are many others (e.g., Ashby, Alfonso-Reese, Turken, & Waldron, 1998; Pothos & Chater, 2002; Rehder & Murphy, 2003; for reviews, see Kruschk”