Prototypes in category learning: the effects of category size, category structure, and stimulus complexity

Although research in categorization has sometimes been motivated by prototype theory, recent studies have favored exemplar theory. However, some of these studies focused on small, poorly differentiated categories composed of simple, 4-dimensional stimuli. Some analyzed the aggregate data of entire groups. Some compared powerful multiplicative exemplar models to less powerful additive prototype models. Here, comparable prototype and exemplar models were fit to individual-participant data in 4 experiments that sampled category sets varying in size, level of category structure, and stimulus complexity (dimensionality). The prototype model always fit the observed data better than the exemplar model did. Prototype-based processes seemed especially relevant when participants learned categories that were larger or contained more complex stimuli.     

Single versus multiple systems of category learning: Reply to Nosofsky and Kruschke (2002)

Nosofsky and Kruschke (2002) argued that the singlesystem ALCOVE model (Kruschke, 1992) can account for the dual-task category learning data reported by Waldron and Ashby (2001). In our reply, we argue that Nosofsky and Kruschke overstated the ability of ALCOVE to account for the Waldron and Ashby results. In fact, ALCOVE has difficulty with these data, and we show that the only versions of ALCOVE that actually fit the Waldron and Ashby accuracy data make incorrect predictions about other previously unreported features of that experiment. We also show that the dual-system COVIS model (Ashby, Alfonso-Reese, Turken, & Waldron, 1998) naturally predicts these results.     

Unsupervised categorization and category learning.

When people categorize a set of items in a certain way they often change their perceptions for these items so that they become more compatible with the learned categorization. In two experiments we examined whether such changes are extensive enough to change the unsupervised categorization for the items-that is, the categorization of the items that is considered more intuitive or natural without any learning. In Experiment 1 we directly employed an unsupervised categorization task; in Experiment 2 we collected similarity ratings for the items and inferred unsupervised categorizations using Pothos and Chater's (2002) model of unsupervised categorization. The unsupervised categorization for the items changed to resemble more the learned one when this was specified by the suppression of a stimulus dimension (both experiments), but less so when it was almost specified by the suppression of a stimulus dimension (Experiment 1, nonsignificant trend in Experiment 2). By contrast, no changes in the unsupervised categorization were observed when participants were taught a classification that was specified by a more fine tuning of the relative salience of the two dimensions.     

A knowledge-resonance (KRES) model of category learning

This article introduces a connectionist model of category learning that takes into account the prior knowledge that people bring to new learning situations. In contrast to connectionist learning models that assume a feedforward network and learn by the delta rule or backpropagation, this model, the knowledge-resonance model, or KRES, employs a recurrent network with bidirectional symmetric connection whose weights are updated according to a contrastive Hebbian learning rule. We demonstrate that when prior knowledge is represented in the network, KRES accounts for a considerable range of empirical results regarding the effects of prior knowledge on category learning, including (1) the accelerated learning that occurs in the presence of knowledge, (2) the better learning in the presence of knowledge of category features that are not related to prior knowledge, (3) the reinterpretation of features with ambiguous interpretations in light of error-corrective feedback, and (4) the unlearning of prior knowledge when that knowledge is inappropriate in the context of a particular category.     

The divergent autoencoder (DIVA) model of category learning

A novel theoretical approach to human category learning is proposed in which categories are represented as coordinated statistical models of the properties of the members. Key elements of the account are learning to recode inputs as task-constrained principle components and evaluating category membership in terms of model fit-that is, the fidelity of the reconstruction after recoding and decoding the stimulus. The approach is implemented as a computational model called DIVA (for DIVergent Autoencoder), an artificial neural network that uses reconstructive learning to solve N-way classification tasks. DIVA shows good qualitative fits to benchmark human learning data and provides a compelling theoretical alternative to established models.     

Implicit and explicit category learning by capuchin monkeys (Cebus apella)

Current theories of human categorization differentiate an explicit, rule-based system of category learning from an implicit system that slowly associates regions of perceptual space with response outputs. The researchers extended this theoretical differentiation to the category learning of New World primates. Four capuchins (Cebus apella) learned categories of circular sine-wave gratings that varied in bar spatial frequency and orientation. The rule-based and information-integration tasks, respectively, had one-dimensional and two-dimensional solutions. Capuchins, like humans, strongly dimensionalized the stimuli and learned the rule-based task more easily. The results strengthen the suggestion that nonhuman primates have some structural components of humans' capacity for explicit categorization, which in humans is linked to declarative cognition and consciousness. The results also strengthen the primate contrast to other vertebrate species that may lack the explicit system. Therefore, the results raise important questions about the origins of the explicit categorization system during cognitive evolution and about its overall phylogenetic distribution.     

Why aggregated learning follows the power law of practice when individual learning does not: comment on Rickard (1997, 1999), Delaney et al. (1998), and Palmeri (1999)

The power law of practice is often considered a benchmark test for theories of cognitive skill acquisition. Recently, P. F. Delaney, L. M. Reder, J. J. Staszewski, and F. E. Ritter (1998), T. J. Palmeri (1999), and T. C. Rickard (1997, 1999) have challenged its validity by showing that empirical data can systematically deviate from power-function fits. The main purpose of the present article is to extend their explanations in two ways. First, the authors empirically show that abrupt changes in performance are not necessarily based on a shift from algorithm to memory-based processing, but rather and more generally, that they occur whenever a more efficient task strategy is generated. Second, the authors show mathematically and per simulation that power functions can perfectly fit aggregated learning curves even when all underlying individual curves are discontinuous. Therefore, the authors question conclusions drawn from fits to aggregated data.     

Single-system models and interference in category learning: Commentary on Waldron and Ashby (2001)

In a recent article, Waldron and Ashby (2001) observed that performing a concurrent task caused greater interference in learning a simple one-dimensional categorization rule than in learning a complex three-dimensional one. They argued that this result was incompatible with all existing single-system models of category learning but was as predicted by the multiple-system COVIS model (Ashby, Alfonso-Reese, Turken, & Waldron, 1998). In contrast to Waldron and Ashby’s argument, we demonstrate that the single-system ALCOVE model (Kruschke, 1992) naturally predicts the result by assuming that its selectiveattention learning process is disrupted by the concurrent task.     

ALCOVE: an exemplar-based connectionist model of category learning.

ALCOVE (attention learning covering map) is a connectionist model of category learning that incorporates an exemplar-based representation (Medin & Schaffer, 1978; Nosofsky, 1986) with error-driven learning (Gluck & Bower, 1988; Rumelhart, Hinton, & Williams, 1986). Alcove selectively attends to relevant stimulus dimensions, is sensitive to correlated dimensions, can account for a form of base-rate neglect, does not suffer catastrophic forgetting, and can exhibit 3-stage (U-shaped) learning of high-frequency exceptions to rules, whereas such effects are not easily accounted for by models using other combinations of representation and learning method.     

From conditioning to category learning: an adaptive network model

We used adaptive network theory to extend the Rescorla-Wagner (1972) least mean squares (LMS) model of associative learning to phenomena of human learning and judgment. In three experiments subjects learned to categorize hypothetical patients with particular symptom patterns as having certain diseases. When one disease is far more likely than another, the model predicts that subjects will substantially overestimate the diagnosticity of the more valid symptom for the rare disease. The results of Experiments 1 and 2 provide clear support for this prediction in contradistinction to predictions from probability matching, exemplar retrieval, or simple prototype learning models. Experiment 3 contrasted the adaptive network model with one predicting pattern-probability matching when patients always had four symptoms (chosen from four opponent pairs) rather than the presence or absence of each of four symptoms, as in Experiment 1. The results again support the Rescorla-Wagner LMS learning rule as embedded within an adaptive network model.     

Meaning and learning: Response to Moerk

In response to Moerk (1986, Developmental Review, 6, 365–385) the following points are discussed: (a) Semantic and conceptual development must be conceived in terms of systems that undergo developmental change, (b) Objects are viewed as embedded in events and as being conceptualized first within a syntagmatic and later a paradigmatic system, (c) While the nature of the input to the child is important, its effects cannot be considered independently of the state of the child's conceptual and linguistic system. (d) The neurological considerations discussed by Moerk are not relevant to the particular developmental issues discussed in my book.     

Combining exemplar-based category representations and connectionist learning rules.

Adaptive network and exemplar-similarity models were compared on their ability to predict category learning and transfer data. An exemplar-based network (Kruschke, 1990a, 1990b, 1992) that combines key aspects of both modeling approaches was also tested. The exemplar-based network incorporates an exemplar-based category representation in which exemplars become associated to categories through the same error-driven, interactive learning rules that are assumed in standard adaptive networks. Experiment 1, which partially replicated and extended the probabilistic classification learning paradigm of Gluck and Bower (1988a), demonstrated the importance of an error-driven learning rule. Experiment 2, which extended the classification learning paradigm of Medin and Schaffer (1978) that discriminated between exemplar and prototype models, demonstrated the importance of an exemplar-based category representation. Only the exemplar-based network accounted for all the major qualitative phenomena; it also achieved good quantitative predictions of the learning and transfer data in both experiments.     

When category learning is holistic: A reply to Ward and Scott

A reply is presented to Ward and Scott’s (1987) recent reservations about the evidence for Kemler Nelson’s (1984) claims about when category learning is likely to be holistic. Focusing on the effect of intention, this paper suggests that: (1) contrary to Ward and Scott’s contention, a reanalysis of a critical set of original data continues to support Kemler Nelson’s claim of more holistic learning under unintentional conditions; (2) there is converging evidence for that claim; (3) Ward and Scott’s incidental learning data may diverge because of the inclusion of many weak learners; (4) Ward and Scott’s counterproposal makes some implausible and unsupported predictions; and (5) some of Ward and Scott’s reaction-time data are difficult to interpret. Still, a final discussion identifies some significant points of agreement with Ward and Scott.     

The neuropsychology of ventral prefrontal cortex: decision-making and reversal learning

Converging evidence from human lesion, animal lesion, and human functional neuroimaging studies implicates overlapping neural circuitry in ventral prefrontal cortex in decision-making and reversal learning. The ascending 5-HT and dopamine neurotransmitter systems have a modulatory role in both processes. There is accumulating evidence that measures of decision-making and reversal learning may be useful as functional markers of ventral prefrontal cortex integrity in psychiatric and neurological disorders. Whilst existing measures of decision-making may have superior sensitivity, reversal learning may offer superior selectivity, particularly within prefrontal cortex. Effective decision-making on existing measures requires the ability to adapt behaviour on the basis of changes in emotional significance, and this may underlie the shared neural circuitry with reversal learning.