Configural and elemental strategies in predictive learning.

Williams, Douglas A.; Sagness, Kerry E.; McPhee, Janice E.

doi:10.1037/0278-7393.20.3.694

Cited by 143 publications

(224 citation statements)

References 45 publications

Supporting

Mentioning

200

Contrasting

Unclassified

Order By: Relevance

“…Our results also parallel a feature of associative models of Pavlovian conditioning and causal learning (see Williams & Braker, 1999), in which experience with compound cues-analogous to our multidimensional objects-results in associations involving those compounds, but not their constituent elements (e.g., Pearce, 1987; see also Shanks, Charles, Darby, & Azmi, 1998;Williams, Sagness, & McPhee, 1994).…”

Section: The Units Of Vslsupporting

confidence: 74%

The Automaticity of Visual Statistical Learning.

Turk‐Browne

Junge

Scholl

2005

Journal of Experimental Psychology: General

673

725

View full text Add to dashboard Cite

Recent studies of visual statistical learning (VSL) have demonstrated that statistical regularities in sequences of visual stimuli can be automatically extracted, even without intent or awareness. Despite much work on this topic, however, several fundamental questions remain about the nature of VSL. In particular, previous experiments have not explored the underlying units over which VSL operates. In a sequence of colored shapes, for example, does VSL operate over each feature dimension independently, or over multidimensional objects in which color and shape are bound together? The studies reported here demonstrate that VSL can be both object-based and feature-based, in systematic ways based on how different feature dimensions covary. For example, when each shape covaried perfectly with a particular color, VSL was object-based: Observers expressed robust VSL for colored-shape sub-sequences at test but failed when the test items consisted of monochromatic shapes or color patches. When shape and color pairs were partially decoupled during learning, however, VSL operated over features: Observers expressed robust VSL when the feature dimensions were tested separately. These results suggest that VSL is object-based, but that sensitivity to feature correlations in multidimensional sequences (possibly another form of VSL) may in turn help define what counts as an object.

show abstract

Section: The Units Of Vslsupporting

confidence: 74%

The Automaticity of Visual Statistical Learning.

Turk‐Browne

Junge

Scholl

2005

Journal of Experimental Psychology: General

673

725

View full text Add to dashboard Cite

show abstract

“…For example, if A signals event O during initial training, and subsequently O is signaled by the combination of A and B, at a final test B on its own will not reliably elicit a response. Blocking is known to play a powerful role in human contingency judgments (e.g., Williams, Sagness, & McPhee, 1994) and, of greatest interest here, in categorization (e.g., Gluck & Bower, 1988;Shanks, 1991Shanks, , 1993). …”

Section: Prior Knowledge and New Learningmentioning

confidence: 99%

Competing strategies in categorization: Expediency and resistance to knowledge restructuring.

Lewandowsky¹,

Kalish²,

Griffiths³

2000

Journal of Experimental Psychology: Learning, Memory, and Cogni

View full text Add to dashboard Cite

The authors investigated people's ability to restructure their knowledge when additional information about a categorization task is revealed. In 2 experiments, people first learned to rely on a fairly accurate (but imperfect) predictor. At various points in training, a complex relationship between 2 other predictors was revealed in a schematic diagram that could support perfect performance. In Experiment 1, people adopted the complex strategy when it was revealed at the outset but were unable to restructure their knowledge after the expedient predictor had been learned. In Experiment 2, expedient knowledge persisted even with an adaptive display. The persistence of expedient knowledge is explained by associative blocking of potential alternative cues. A 3rd experiment analyzed the strategies people use with and without the diagram. The study confirmed that the diagram, when presented at the outset, significantly alters people's approach to the task.Knowledge restructuring is often observed when people develop new skills, both in the real world and in the laboratory; paradoxically, expertise at these skills in some cases also seems to prevent restructuring of knowledge. Little is known about the processes involved in knowledge restructuring, and our investigation formed a first step toward describing those processes in order to better understand when restructuring occurs and when it is resisted.Specifically, we examined people's ability to adapt and restructure their knowledge when previously concealed information about a categorization task was revealed halfway through training. The task could be performed either by relying on an expedient single predictor that permitted fairly accurate (but imperfect) performance or by learning one of several complex (but potentially perfect) strategies involving two predictors. It was known from previous research that people can use a complex strategy for this task if the relationship between predictors is revealed in a schematic diagram before training (Lewandowsky, Dunn, Kirsner, & Randell, 1997). Here, we asked two questions: First, can people use the diagram to restructure their knowledge after having learned to rely on the expedient predictor, and what are the processes underlying that restructuring? Identification of those processes requires exploration of the second question, namely what exact strategy or rule people used with and without the diagram. We report three experiments in which the diagram was presented at various points in training.Stephan Lewandowsky, Mike Kalish, and Thomas L. Griffiths, Department of Psychology, University of Western Australia, Nedlands, Australia.Preparation of this article was facilitated by Large Research Grants A79600016 and A79941108 from the Australian Research Council. We thank Janet Phang and Lih Yi Soong for assistance during data collection and analysis, and Mike Mundy and Christina Chandler for help during preparation of this article.Correspondence concerning this article should be addressed to Stephan Lewandowsky, Department of...

show abstract

“…For example, studies investigating learning of stimulus-outcome relations in human subjects have shown that performance to an overshadowed stimulus can be retroactively increased by posttraining exposures to the paired stimulus. Specifically, if subjects learn the association between a compound cue, AB, and a particular outcome, then the predictive value accorded A can be retroactively decreased by paired presentations of B and the outcome (Chapman, 1991;Dickinson & Burke, 1996;Shanks, 1985;Van Hamme & Wasserman, 1994;Williams, Sagness, & McPhee, 1994) or retroThe research was supported by a grant from the Australian Research Council, and the experimental procedures were approved by the Animal Care and Ethics Committee (ACEC) of the University of New South Wales, The authors are grateful to P. Lovibond and R. Richardson for comments. Correspondence should be addressed to 1.A.…”

mentioning

confidence: 99%

Retroactive revaluation of an odor-taste association

Harris

Westbrook

1998

Animal Learning & Behavior

View full text Add to dashboard Cite

Rats showed an increase or decrease in preference for an aqueous odor conditioned stimulus (CS) presented in compound with a palatable (sucrose) or unpalatable (quinine) taste unconditioned stimulus (US), respectively. Four experiments then studied the consequences for odor preference of presenting water in the context where the odor-taste compound had previously occurred (i.e., no-CSno-US experience). Exposure to that context increased the preference for the odor associate of sucrose and decreased preference for the odor associate of quinine. The increased preference for the odor associate of sucrose was contingent on reexposure to the context where the odor-sucrose compound had occurred (i.e., retroactive revaluation) and was reduced by preexposure to the context where the odor-sucrose compound was to occur (Le.,latent inhibition of the context). The retroactive revaluation of within-event learning caused by exposures to the associated context is discussed in terms of (1) the modified version of the SOP model (Wagner, 1981) proposed by Dickinson and Burke (1996)and (2) the comparator model (Miller & Matzel, 1988).Contemporary theories ofassociative learning have been strongly influenced by demonstrations that responding to a target conditioned stimulus (CS) is modulated by the status of that and other CSs present during training. Interference effects such as overshadowing, blocking, and latent inhibition have inspired theories that describe the mechanisms by which CSs compete for attentional processing and for associative strength with the unconditioned stimulus (US) (Mackintosh, 1975;Pearce & Hall, 1980;Rescorla & Wagner, 1972;Wagner, 1981). Thus, these theories typically attribute variations in conditioned responding to differences in the acquisition ofassociations during CS-US pairings. Moreover, these theories assume that modulation of CS-US associations can occur only when the animal is directly exposed to the CS.However, this view has been challenged by demonstrations that performance to a CS can be increased or decreased retroactively by changing the associative strength between other CSs and the US. For example, studies investigating learning of stimulus-outcome relations in human subjects have shown that performance to an overshadowed stimulus can be retroactively increased by posttraining exposures to the paired stimulus. Specifically, if subjects learn the association between a compound cue, AB, and a particular outcome, then the predictive value accorded A can be retroactively decreased by paired presentations of B and the outcome (Chapman, 1991;Dickinson

show abstract

Configural and elemental strategies in predictive learning.

Cited by 143 publications

References 45 publications

The Automaticity of Visual Statistical Learning.

The Automaticity of Visual Statistical Learning.

Competing strategies in categorization: Expediency and resistance to knowledge restructuring.

Retroactive revaluation of an odor-taste association

Contact Info

Product

Resources

About