An adaptive linear filter model of procedural category learning

Marchant, Nicolás; Canessa, Enrique; Chaigneau, Sergio E.

doi:10.1007/s10339-022-01094-1

Cited by 4 publications

(3 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In its current state, the ALF model assumes only one free parameter (i.e., the learning rate), and one can assume a second free parameter used for scaling in the logistic function and representing the deterministic course of learning over the experiment (i.e., controlling the curve’s slope). In fact, elsewhere we have shown that successful modeling of human category learning data using the ALF model is possible (Marchant et al, 2022).…”

Section: Discussionmentioning

confidence: 99%

“…The simulations that will be presented in the next sections were performed using the ALF model (Marchant et al, 2022). Similar to the models originally implemented by Gluck and Bower (1988a, 1988b), the ALF we discuss here assumes that organisms update coefficients for each discrete feature, such that classification errors relative to feedback are minimized.…”

Section: Adaptive Linear Filter Model and Category Learning Setupsmentioning

confidence: 99%

See 1 more Smart Citation

On the importance of feedback for categorization: Revisiting category learning experiments using an adaptive filter model.

Marchant¹,

Chaigneau²

2022

Journal of Experimental Psychology: Animal Learning and Cogniti

Self Cite

View full text Add to dashboard Cite

Associative accounts of category learning have been, for the most part, abandoned in favor of cognitive explanations (e.g., similarity, explicit rules). In the current work, we implement an Adaptive Linear Filter (ALF) closely related to the Rescorla and Wagner learning rule, and use it to tackle three learning tasks that pose challenges to an associative view of category learning. Across three computational simulations, we show that the ALF is in fact able to make the predictions that seemed problematic. Notably, in our simulations we use exactly the same model and specifications, attesting to the generality of our account. We discuss the consequences of our findings for the category learning literature.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Adaptive Linear Filter Model and Category Learning Setupsmentioning

confidence: 99%

On the importance of feedback for categorization: Revisiting category learning experiments using an adaptive filter model.

Marchant¹,

Chaigneau²

2022

Journal of Experimental Psychology: Animal Learning and Cogniti

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus, they might have not considered the category membership rating question as a causal‐probabilistic problem. For example, they could have conceptualized the individual features and the causal relation as three independent variables—with the causal relation becoming a configural cue—combined in an associative computation (e.g., Gluck & Bower, 1988; Gluck & Myers, 2001, Marchant, Canessa, & Chaigneau, 2022). This would account for the lower weight of causal information and higher relative weight of featural information in the category membership condition.…”

Section: Methodsmentioning

confidence: 99%

A Context‐Dependent Bayesian Account for Causal‐Based Categorization

2023

Self Cite

View full text Add to dashboard Cite

The causal view of categories assumes that categories are represented by features and their causal relations. To study the effect of causal knowledge on categorization, researchers have used Bayesian causal models. Within that framework, categorization may be viewed as dependent on a likelihood computation (i.e., the likelihood of an exemplar with a certain combination of features, given the category's causal model) or as a posterior computation (i.e., the probability that the exemplar belongs to the category, given its features). Across three experiments, in combination with computational modeling, we offer evidence that categorization is better accounted for by assuming that people compute posteriors and not likelihoods, though both probabilities are closely related. This result contrasts with existing analyses of causal-based categorization, which assume that likelihood computations give a good approximation of human judgments. We also find that people are able to compute likelihoods in a closely related task that elicits judgments of consistency rather than category membership judgments. Our analyses show that people do use causal probabilistic information as prescribed by a Bayesian model but that they flexibly compute likelihoods or posteriors depending on the task. We discuss our results in relation to the relevant literature on the topic.

show abstract