The Wisdom of Model Crowds

He, Lisheng; Analytis, Pantelis P.; Bhatia, Sudeep

doi:10.1287/mnsc.2021.4090

Cited by 15 publications

(18 citation statements)

References 113 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(Note, however, that that this result is not trivial, as the penalty incurred by Models 2-4 for including additional parameters could have caused them to lose the comparison.) Indeed, previous studies have shown (using large-scale datasets) that EV alone is not the best model for choice (He et al, 2022;Peterson et al, 2021). To increase the ability of the integrate-then-compare model (Model 1) to explain the data, we considered a variant where in the place of EV for each option we calculated subjective utility as under Prospect Theory (Tversky & Kahneman, 1992).…”

Section: Model Comparisonmentioning

confidence: 99%

Risky decisions are influenced by individual attributes as a function of risk preference

Lee¹,

D’Alessandro²,

Iodice³

et al. 2022

Preprint

View full text Add to dashboard Cite

It has long been assumed in economic theory that multi-attribute decisions involving several attributes or dimensions – such as probabilities and amounts of money to be earned during risky choices – are resolved by first combining the attributes of each option to form an overall expected value and then comparing the expected values of the alternative options, using a unique evidence accumulation process (integrate-then-compare model). Two plausible alternatives would be: performing multiple comparisons between the individual attributes in parallel and then integrating the results of the comparisons afterwards (compare-then-integrate model); combining the two methods described above (combined model). Distinguishing between these alternative models has been difficult, as they tend to generate similar predictions across a wide range of conditions. Here, we devise a novel method to disambiguate between integrate-then-compare, compare-then-integrate, and combined models of multi-attribute decisions, by orthogonally manipulating the expected value of the options and the saliency of their attributes. Our results, using behavioral measures and drift-diffusion models, provide evidence in favor of the compare-then-integrate and combined models but against the integrate-then-compare model. This suggests that risky decisions are resolved by running in parallel multiple comparisons between the separate attributes – possibly in combination with an additional comparison between the expected values of the options. This result stands in contrast with the assumption of standard economic theory that choices require a unique comparison of expected values and suggests that at the cognitive and neural levels, decision processes might be more distributed than commonly assumed.

show abstract

Section: Model Comparisonmentioning

confidence: 99%

Risky decisions are influenced by individual attributes as a function of risk preference

Lee¹,

D’Alessandro²,

Iodice³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…An alternative view involves the notion of models as collaborative crowds rather than as fierce competitors. He et al (2021) suggested that when there is no unequivocal best-performing model for individual findings, the aggregation of models can flexibly account for larger shares of data using the "wisdom of crowds." Accordingly, this exploits the unique complementarities of different models, bringing the research field closer to reaching its full potential.…”

Section: Concluding Remarks and Future Directionsmentioning

confidence: 99%

Conditioned inhibition, inhibitory learning, response inhibition, and inhibitory control: Outlining a conceptual clarification

Sosa¹

2022

Preprint

View full text Add to dashboard Cite

Inhibition can be defined as a phenomenon in which an agent prevents or suppresses a behavioral state that would otherwise occur. Associative learning studies have extensively examined how experiences shape the acquisition of inhibitory behavioral tendencies across many species and situations. Associative inhibitory phenomena can be studied at various levels of analysis. One could focus on the trajectory of behavioral change involved in learning from negative statistical associations between discrete events (inhibitory learning). Alternatively, one could be interested in the effects of accumulated experience with those negative associations (conditioned inhibition). One could rather be interested in how organisms implement what they learn through experiences involving negative associations (response inhibition). Yet one could inquire into how the capacity of learning negative associations and performing accordingly varies between individuals and along time for the same individual (inhibitory control). This paper presents a tentative taxonomy addressing different levels of analysis of associative inhibitory phenomena by using different terms for each. In addition, recent evidence and certain unresolved issues at each level are thoroughly scrutinized and contrasted with prior findings. The empirical and theoretical advances made by modeling inhibition as an associative learning phenomenon have provided scaffolds for the current knowledge and emerging accounts of the topic. Some of those emerging accounts have the potential to bridge different levels of analysis and foster “cross-pollination” of ideas among broad fields beyond associative learning.

show abstract

“…However, this model is limited in terms of scalablity and adaptability. For example, it cannot be easily trained and provide predictions for new data (Peterson et al, 2021;He et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

AS-BEAST: Using machine learning to create an adaptable, scalable, and interpretable behavioral model

Shoshan¹,

Hazan²,

Plonsky³

2023

Preprint

View full text Add to dashboard Cite

In this paper, we propose a behavioral model called BEAST-Net, which combines the basic logic of BEAST, a psychological theory-based behavioral model, with machine learning (ML) techniques. Our approach is to formalize BEAST mathematically as a differentiable function and parameterize it with a neural network, enabling us to learn the model parameters from data and optimize it using backpropagation. The resulting model, BEAST-Net, is able to scale to larger datasets and adapt to new data with greater ease, while retaining the psychological insights and interpretability of the original model. We evaluate BEAST-Net on the largest public benchmark dataset of human choice tasks and show that it outperforms several baselines, including the original BEAST model. Furthermore, we demonstrate that our model can be used to provide interpretable explanations for choice behavior, allowing us to derive new psychological insights from the data. Our work makes a significant contribution to the field of human decision making by showing that ML techniques can be used to improve the scalability and adaptability of psychological theory based models while preserving their interpretability and ability to provide insights.

show abstract

The Wisdom of Model Crowds

Cited by 15 publications

References 113 publications

Risky decisions are influenced by individual attributes as a function of risk preference

Risky decisions are influenced by individual attributes as a function of risk preference

Conditioned inhibition, inhibitory learning, response inhibition, and inhibitory control: Outlining a conceptual clarification

AS-BEAST: Using machine learning to create an adaptable, scalable, and interpretable behavioral model

Contact Info

Product

Resources

About