Similar task features shape judgment and categorization processes.

Hoffmann, Janina A.; Helversen, Bettina von; Rieskamp, Jörg

doi:10.1037/xlm0000241

Cited by 24 publications

(68 citation statements)

References 73 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Beyond that, our findings more closely followed the simulations with a Radial Basis Function kernel than what would have been expected if participants had extrapolated more linearly (e.g. Busemeyer et al, 1997;Hoffmann, von Helversen, & Rieskamp, 2016). Using an RBF kernel, our model predicted a relatively small difference in choice proportions between the exotic-novel and ordinary-novel options, but the difference could have been larger if extrapolation relied on a different kernel which does not underestimate the average reward of the exotic-novel option.…”

Section: Discussionsupporting

confidence: 85%

It’s new, but is it good? How generalization and uncertainty guide the exploration of novel options.

Stojić¹,

Schulz²,

Analytis³

et al. 2020

Journal of Experimental Psychology: General

View full text Add to dashboard Cite

How do people decide whether to try out novel options as opposed to tried-and-tested ones? We argue that they infer a novel option's reward from contextual information learned from functional relations and take uncertainty into account when making a decision. We propose a Bayesian optimization model to describe their learning and decision making. This model relies on similarity-based learning of functional relationships between features and rewards, and a choice rule that balances exploration and exploitation by combining predicted rewards and the uncertainty of these predictions. Our model makes two main predictions. First, decision makers who learn functional relationships will generalize based on the learned reward function, choosing novel options only if their predicted reward is high. Second, they will take uncertainty about the function into account, and prefer novel options that can reduce this uncertainty. We test these predictions in three preregistered experiments in which we examine participants' preferences for novel options using a feature-based multi-armed bandit task in which rewards are a noisy function of observable features. Our results reveal strong evidence for functional exploration and moderate evidence for uncertainty-guided exploration. However, whether or not participants chose a novel option also depended on their attention, as well as reflecting on the value of the options. These results advance our understanding of people's reactions in the face of novelty.

show abstract

Section: Discussionsupporting

confidence: 85%

It’s new, but is it good? How generalization and uncertainty guide the exploration of novel options.

Stojić¹,

Schulz²,

Analytis³

et al. 2020

Journal of Experimental Psychology: General

View full text Add to dashboard Cite

show abstract

“…We found that the best overall model was a linear regression combined with an UCB sampling strategy. This means that participants adapted well to the overall task structure, which was set up to be linear, a finding which is in line with previous work showing that participants rely more on rule-based than similarity-based reasoning in simple linear environments (Hoffmann, von Helversen, & Rieskamp, 2016). Moreover, they sampled inputs to learn about both the overall shape of the function as well as the location of high outputs.…”

Section: Discussionsupporting

confidence: 83%

Active Function Learning

Jones

Schulz²,

Meder

et al. 2018

Preprint

View full text Add to dashboard Cite

How do people actively explore to learn about functional rules, that is, how continuous inputs map onto continuous outputs? We introduce a novel paradigm to investigate information search in continuous, multi-feature function learning scenarios. Participants either actively selected or passively observed information to learn about an underlying linear function. We develop and compare different variants of rule-based (linear regression) and non-parametric (Gaussian process regression) active learning approaches to model participants' active learning behavior. Our results show that participants' performance is best described by a rule-based model that attempts to efficiently learn linear functions with a focus on high and uncertain outcomes. These results advance our understanding of how people actively search for information to learn about functional relations in the environment.

show abstract

“…In the nonlinear environment, they tend to adopt a similarity based strategy, predicting values based on how similar the item at hand is to previously experienced items stored in memory (e.g. Hoffmann et al, 2016;Ashby & Maddox, 2011). Participants first underwent a training phase in which they were given the true function values, y, as feedback.…”

Section: Methodsmentioning

confidence: 99%

“…Nosofsky, 1984), and the cognitive processes behind them remain mysterious. Influential models of both category learning (concepts with discrete predicted variables) and function learning (respectively continuous; which we study here), only model predictions and thus cannot easily account for confidence judgements (Hoffmann, von Helversen, & Rieskamp, 2016;Mc-Daniel & Busemeyer, 2005). In contrast, confidence has been extensively studied in perceptual decision making and sensorimotor learning (Pleskac & Busemeyer, 2010;Moran, Teodorescu, & Usher, 2015;Körding & Wolpert, 2004).…”

Section: Introductionmentioning

confidence: 99%

“…We varied the type of environment within participants -including linear and quadratic mappings between item features and values. This allowed us address the qualitatively different learning strategies implied in past literature: rule-or similarity-based strategies, that are induced by these environments (Hoffmann et al, 2016;Ashby & Maddox, 2011). We also varied between participants whether they provided predictions ("Pred only" condition) or both predictions and confidence ratings ("Pred CR"), to examine whether making confidence judgements itself changes the learning process 1 .…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Are you sure about that? On the origins of confidence in concept learning

Stojić

Eldar

Bassam

et al. 2018

2018 Conference on Cognitive Computational Neuroscience

View full text Add to dashboard Cite

Humans possess a rich repertoire of abstract concepts about which they can often judge their confidence. These judgements help guide behaviour, but the mechanisms underlying them are still poorly understood. Here, we examine the evolution of people's sense of confidence as they engage in probabilistic concept learning. Participants learned a continuous function of four continuous features, reporting their predictions and confidence about these predictions. Participants indeed had insight into their uncertainties: confidence was correlated with the accuracy of predictions, increasing as learning progressed. There were substantial individual differences. In contrast to many classical models that try to explain only the predictions, we formalized human function learning in Bayesian terms as Gaussian process inference. This model generates posterior distributions, allowing us to link predictions and confidence judgements. Gaussian process inference well matched participants' predictions, and also the confidence judgements of metacognitively competent participants. Our results show that human confidence judgements during learning are tied to uncertainty, suggesting that concept learning is fundamentally probabilistic.

show abstract

Similar task features shape judgment and categorization processes.

Cited by 24 publications

References 73 publications

It’s new, but is it good? How generalization and uncertainty guide the exploration of novel options.

It’s new, but is it good? How generalization and uncertainty guide the exploration of novel options.

Active Function Learning

Are you sure about that? On the origins of confidence in concept learning

Contact Info

Product

Resources

About