Rule-based extrapolation in perceptual categorization

Erickson and Kruschke (1998) provided a demonstration that in certain situations people will classify novel stimuli according to an extrapolated rule, even when the most similar training exemplar is an exception to the rule. This result challenged exemplar models. Nosofsky and Johansen (2000) have called this finding into question by offering an exemplar-based explanation for those data based on the perceptual features of the stimuli. Here, we describe the results of a new experiment that yields results similar to those found previouslywithout the questionable perceptual features:Participantswho learn to classify all the training stimuli have patterns of generalization that indicate a combination of rule and exemplar representation. ATRIUM, a hybrid rule and exemplar model (Erickson & Kruschke, 1998), is shown to account for these data much better than ALCOVE, an exemplar model (Kruschke, 1992).Moreover, four alternate exemplar explanations, including one suggested by Nosofsky and Johansen, cannot account for our new findings.     

知道与不知道要学多少：类别学习中样例量的预期作用

采用“5/4模型”类别结构探讨了类别学习中样例量的预期作用。设置了两种学习条件(“知道样例量”和“不知道样例量”), 分别探讨两种学习条件下的学习效率、学习策略以及所形成的类别表征。106名大学生参加了实验, 结果表明：在类别学习中, 样例量的预期作用显著, 知道样例量组的学习效率高于不知道样例量组; 样例量的预期作用对类别学习效率的影响是通过影响学习过程中使用的策略来实现的; 样例量的预期作用不影响两种学习条件的学习后形成的类别表征, 且两种学习条件的被试自始至终表现出样例学习的表征模式。     

Comparing decision bound and exemplar models of categorization

The performance of a decision bound model of categorization (Ashby, J992a; Ashby & Maddox, in press) is compared with the performance of two exemplar models. The first is the generalized context model (e.g., Nosofsky, 1986, 1992) and the second is a recently proposed deterministic exemplar model (Ashby & Maddox, in press), which contains the generalized context model as a special case. When the exemplars from each category were normally distributed and the optimal decision bound was linear, the deterministic exemplar model and the decision bound model provided roughly equivalent accounts of the data. When the optimal decision bound was nonlinear, the decision bound model provided a more accurate account of the data than did either exemplar model. When applied to categorization data collected by Nosofsky (1986, 1989), in which the category exemplars are not normally distributed, the decision bound model provided excellent accounts of the data, in many cases significantly outperforming the exemplar models. The decision bound model was found to be especially successful when(1) single subject analyses were performed, (2) each subject was given relatively extensive training, and (3) the subject's performance was characterized by complex suboptimalities. These results support the hypothesis that the decision bound is of fundamental importance in predicting asymptotic categorization performance and that the decision bound models provide a viable alternative to the currently popular exemplar models of categorization.     

Contrast and assimilation in categorization and exemplar production

We studied contrast and assimilation in three tasks: an exemplar-production task, a categorization task, and a combined categorization-then-production task. On each trial of the first task, subjects produced a circle when prompted with a category label. In the second task, they classified lines that differed in length into one of four categories. On each trial of the combined task, they classified two lines and then produced a line when prompted by a category label. All three tasks converged on the same conclusion: subjects' representation of the categories (measured in pixels in the production tasks and by the direction of errors in classification) shifted systematically from trial to trial. When successive stimuli were from the same category, the representation of that category was pulled toward the exemplar from the previous trial. When successive stimuli were from different categories, the representations of the neighbouring categories were pushed from the category of the initial stimulus. We conclude that accounts of categorization and identification must accommodate both assimilation and contrast as a function of trial-to-trial shifts in representation.     

Exemplars and prototypes in natural language concepts: A typicality-based evaluation

Are natural language categories represented by instances of the category or by a summary representation? We used an exemplar model and a prototype model, both derived within the framework of the generalized context model (Nosofsky, 1984, 1986), to predict typicality ratings for 12 superordinate natural language concepts. The models were fitted to typicality ratings averaged across participants and to the typicality judgments of individual participants. Both analyses yielded results in favor of the exemplar model. These results suggest that higher-level natural language concepts are represented by their subordinate members, rather than by a summary representation.     

Models for letter cancellation performance and their implications for models of reading

Models for letter cancellation are developed, i.e. for tasks where subjects read a coherent piece of prose at the same time as trying to cancel every target letter they encounter. These models have two components (i) a model of how information is represented by the reader, (ii) a model of how this information is scanned to enable the reader to detect particular target letters. Models of representation range from a photographic image of the printed page to a sophisticated multi-level representation, and models of scanning range from a series of independent saccadic movements (like eye movements) to scans which take into account morphemic and syntactic structure.

Several predictions about the detailed pattern of letter cancellation data are developed on the basis of these models, and these are tested on a data base collected by Smith and Groat (1979). While some support for the simpler models is forthcoming, only the most elaborate models (multi-level representation, syntactically guided scan) can fully account for all the data. While instances can be found where the cancellation task is clearly disrupting the reading task, much of the data are consistent with the notion that the letter cancellation task is not excessively intrusive, and our analyses can thus give us significant information about normal reading.     

The integrative action of the cerebral hemispheres

In four experiments, right-handed subjects were asked to identify uppercase letters presented tachistoscopically in the right (R) or left (L) visual field or in both visual fields simultaneously (RL). When R, L, and RL trials were randomized together and RL trials consisted of the same letter in each visual field (Experiments 1 and 2), 11 of 21 male subjects (Experiment 1) and 9 of 18 female subjects (Experiment 2) showed a strong left-visual-field advantage, and the remainder in each experiment showed an equivalently strong right-visual-field advantage. When the RL trials consisted of a different letter in the two visual fields (Experiment 3), a consistent right-visual-field advantage was observed. It is argued that these results reflect predominantly analytical (left-hemisphere) processing of “different” pairs and relatively holistic processing of “same” pairs, which induces a shift toward a right-hemisphere advantage in some subjects. The main purpose of three of the four experiments was to test five probability models of hemispheric integration: (1) statistical summation, which assumes that the hemispheres operate independently; (2) redundancy, which assumes that RL decisions reflect only processing by the more specialized hemisphere; (3) complete dependency, according to which RL decisions are the mean of Rand L decisions; (4) integration, which assumes that R and L decisions can be represented as vectors in a joint RL space; (5) correlation, according to which the Rand L decisions are assumed to be statistically dependent. Whether R, L, and RL trials were randomized together or RL trials were presented in separate blocks, models 1 and 4 could be clearly rejected. In general, the best predictor of RL performance was model 5. It is argued that the hemispheres function as an integrated system in letter identification.     

Not just the norm: Exemplar-based models also predict face aftereffects

The face recognition literature has considered two competing accounts of how faces are represented within the visual system: Exemplar-based models assume that faces are represented via their similarity to exemplars of previously experienced faces, while norm-based models assume that faces are represented with respect to their deviation from an average face, or norm. Face identity aftereffects have been taken as compelling evidence in favor of a norm-based account over an exemplar-based account. After a relatively brief period of adaptation to an adaptor face, the perceived identity of a test face is shifted toward a face with attributes opposite to those of the adaptor, suggesting an explicit psychological representation of the norm. Surprisingly, despite near universal recognition that face identity aftereffects imply norm-based coding, there have been no published attempts to simulate the predictions of norm- and exemplar-based models in face adaptation paradigms. Here, we implemented and tested variations of norm and exemplar models. Contrary to common claims, our simulations revealed that both an exemplar-based model and a version of a two-pool norm-based model, but not a traditional norm-based model, predict face identity aftereffects following face adaptation.     

Idealness and similarity in goal-derived categories: A computational examination

The finding that the typicality gradient in goal-derived categories is mainly driven by ideals rather than by exemplar similarity has stood uncontested for nearly three decades. Due to the rather rigid earlier implementations of similarity, a key question has remained—that is, whether a more flexible approach to similarity would alter the conclusions. In the present study, we evaluated whether a similarity-based approach that allows for dimensional weighting could account for findings in goal-derived categories. To this end, we compared a computational model of exemplar similarity (the generalized context model; Nosofsky, Journal of Experimental Psychology. General 115:39–57, 1986) and a computational model of ideal representation (the ideal-dimension model; Voorspoels, Vanpaemel, & Storms, Psychonomic Bulletin & Review 18:1006-114, 2011) in their accounts of exemplar typicality in ten goal-derived categories. In terms of both goodness-of-fit and generalizability, we found strong evidence for an ideal approach in nearly all categories. We conclude that focusing on a limited set of features is necessary but not sufficient to account for the observed typicality gradient. A second aspect of ideal representations—that is, that extreme rather than common, central-tendency values drive typicality—seems to be crucial.     

The memory representation of motor skills: a test of schema theory

Currently, a popular model for the central representation of motor skills is embodied in Schmidt's schema theory of discrete motor skill learning (Schmidt, 1975). Two experiments are reported here that contrast predictions from a schema abstraction model that is the basis for schema theory with those from an exemplar-based model of motor skill memory representation. In both experiments, subjects performed 300 trials per day of three variations of a three-segment timing task over 4 days of acquisition. The subjects then either immediately transferred to four novel variations of the same task (Experiment 1) that varied in degree of similarity to the exemplars experienced during acquisition; or performed two novel and two previously produced exemplars, following 24-h and 1-week retention intervals (Experiment 2). The results indicated that novel task transfer was not affected by the degree of similarity between the acquisition and transfer exemplars, and that there was no advantage for a previously produced exemplar over a novel exemplar after either a 24-hr or 1 week retention interval. Also, in both experiments, a consistent pattern of bias in responding was noted for novel task transfer and retention. These results are indicative of a schema abstraction model of memory representation for motor skills.     

When the Rule is Ruled Out: Exemplars and Rules in Decisions from Memory

What are the cognitive processes underlying people's decisions from memory? Previous research suggests that these processes can be best described by strategies that are based on abstract knowledge about the decision task (e.g., cue–criterion relations). However, recent results show that different cue presentation formats trigger the use of different decision strategies and that exemplar‐based strategies can account for memory‐based decisions as well. In three experiments, this effect was replicated, and mediation analyses identified the quality of the underlying knowledge representation as the critical variable that determines decision behavior. This supports the view that exemplar‐based reasoning is used as a backup system if cue abstraction is hindered. Copyright © 2012 John Wiley & Sons, Ltd.     

On coding the position of letters in words: a test of two models

Open-bigram and spatial-coding schemes provide different accounts of how letter position is encoded by the brain during visual word recognition. Open-bigram coding involves an explicit representation of order based on letter pairs, while spatial coding involves a comparison function operating over representations of individual letters. We identify a set of priming conditions (subset primes and reversed interior primes) for which the two types of coding schemes give opposing predictions, hence providing the opportunity for strong scientific inference. Experimental results are consistent with the open-bigram account, and inconsistent with the spatial-coding scheme.     

Category representation for classification and feature inference

This research's purpose was to contrast the representations resulting from learning of the same categories by either classifying instances or inferring instance features. Prior inference learning research, particularly T. Yamauchi and A. B. Markman (1998), has suggested that feature inference learning fosters prototype representation, whereas classification learning encourages exemplar representation. Experiment 1 supported this hypothesis. Averaged and individual participant data from transfer after inference training were better fit by a prototype than by an exemplar model. However, Experiment 2, with contrasting inference learning conditions, indicated that the prototype model was mimicking a set of label-based bidirectional rules, as determined by the inference learning task demands in Experiment 1. Only the set of rules model accounted for all the inference learning conditions in these experiments.     

The effects of visual similarity on proofreading for misspellings

In three experiments, subjects read prose passages and circled misspellings in them. The misspelled words were created by replacing a single letter with another one. The visual similarity between the correct letter and the one that was substituted for it increased the percentage of proofreading errors. The results suggest that proofreaders search through a visual representation of the text and that a hierarchical feature test is applied to this representation, according to which subjects give first priority to resolving letter envelope and second priority to discriminating additional visual features. A sophisticated-guessing decision rule is also implicated for the misspellings that do not alter letter envelope: Subjects are tolerant of missing letter features but are intolerant of additional features.     

Journey to the center of the category: the dissociation in amnesia between categorization and recognition

The authors' theoretical analysis of the dissociation in amnesia between categorization and recognition suggests these conclusions: (a) Comparing to-be-categorized items to a category center or prototype produces strong prototype advantages and steep typicality gradients, whereas comparing to-be-categorized items to the training exemplars that surround the prototype produces weak prototype advantages and flat typicality gradients; (b) participants often show the former pattern, suggesting their use of prototypes; (c) exemplar models account poorly for these categorization data, but prototype models account well for them; and (d) the recognition data suggest that controls use a single-comparison exemplar-memorization process more powerfully than amnesics. By pairing categorization based in prototypes with recognition based in exemplar memorization, the authors support and extend other recent accounts of cognitive performance that intermix prototypes and exemplars, and the authors reinforce traditional interpretations of the categorization-recognition dissociation in amnesia.     

Representing word meaning and order information in a composite holographic lexicon

The authors present a computational model that builds a holographic lexicon representing both word meaning and word order from unsupervised experience with natural language. The model uses simple convolution and superposition mechanisms (cf. B. B. Murdock, 1982) to learn distributed holographic representations for words. The structure of the resulting lexicon can account for empirical data from classic experiments studying semantic typicality, categorization, priming, and semantic constraint in sentence completions. Furthermore, order information can be retrieved from the holographic representations, allowing the model to account for limited word transitions without the need for built-in transition rules. The model demonstrates that a broad range of psychological data can be accounted for directly from the structure of lexical representations learned in this way, without the need for complexity to be built into either the processing mechanisms or the representations. The holographic representations are an appropriate knowledge representation to be used by higher order models of language comprehension, relieving the complexity required at the higher level.     

Hemisphere-specific effects in word recognition do not require hemisphere-specific modes of access

Consistent with converging experimental evidence, we assume that foveal information is initially split across the two cerebral hemispheres. We have previously presented the SERIOL model of letter-position coding, which specifies how the resulting two halves of a letter string are integrated into an abstract representation of letter order. This representation is based on ordered pairs of two letters (bigrams); such a representation is created for input occurring at any location in the visual field. Various studies have shown hemisphere-specific effects in lexical access by using unilateral presentation of stimuli. While the hemisphere-independent means of lexical access in the SERIOL model (via bigrams) may seem inconsistent with these findings, we propose that such effects arise from the hemisphere-specific transformations necessary to create the bigram representation. We provide a theoretical account of the finding that high N (lexical neighborhood size) evokes facilitation in the RH, but not the LH (Lavidor & Ellis, 2002a, 2002b); an experiment discussed elsewhere (Whitney & Lavidor, 2003) has verified key predictions of this account. We also discuss the differing effects of word length across visual fields.     

Fast visual generation: its nature and chronometrics.

The fast-generation model for the matching of mixed-case letter pairs (e.g., Aa, Ab) states that one or both members of a pair activate visual representations in memory of the opposite case, supporting "same" or "different" responses through crossmatching to representations of the pair members themselves. Here the reaction time and error results of three experiments using simultaneous matches support a specific variant of the model in which generation proceeds from the uppercase letter. Furthermore, a manipulation of stimulus onset asynchrony in a fourth experiment using near-simultaneous matches indicates that fast generation produces a visual representation that occurs within 67 msec of initiation and that decays within 200 msec. A fifth experiment contrasts simultaneous and successive matches and in the case of successive matches finds evidence in support of a regeneration process acting after an initial decay. Models of mixed-case matching that are based on the phonetic representation of letter names, or on abstract-letter identities, completely fail to account for the results. Fast generation is distinguishable from slow generation in that it shows fast (vs. slow) dynamics, rapid decay (vs. maintainability), no imagery (vs. imagery), and (probably) automatic (vs. controlled) processing.     

How many exemplars are used? Explorations with the Rex Leopold I model

The goal of this research is to test the hypothesis that a category is not necessarily represented by all observed exemplars, but by a reduced subset of these exemplars. To test this hypothesis, we made use of a study reported by Nosofsky, Clark, and Shin (1989), and replicated their Experiment 1 in order to gather individual-participant data. Both a full exemplar model and a reduced exemplar model were fit to the data. In general, the fits of the reduced exemplar model were superior to those of the full exemplar model. The results suggest that only a subset of exemplars may be sufficient for category representation.     

Fast visual generation: Its nature and chronometrics

The fast-generation model for the matching of mixed-case letter pairs (e.g., Aa, Ab) states that one or both members of a pair activate visual representations in memory of the opposite case, supporting “same” or “different” responses through crossmatching to representations of the pair members themselves. Here the reaction time and error results of three experiments using simultaneous matches support a specific variant of the model in which generation proceeds from the uppercase letter. Furthermore, a manipulation of stimulus onset asynchrony in a fourth experiment using near-simultaneous matches indicates that fast generation produces a visual representation that occurs within 67 msec of initiation and that decays within 200 msec. A fifth experiment contrasts simultaneous and successive matches and in the case of successive matches finds evidence in support of a regeneration process acting after an initial decay. Models of mixed-case matching that are based on the phonetic representation of letter names, or on abstract-letter identities, completely fail to account for the results. Fast generation is distinguishable from slow generation in that it shows fast (vs. slow) dynamics, rapid decay (vs. maintainability), no imagery (vs. imagery), and (probably) automatic (vs. controlled) processing.