A Boundedly Rational Model for Category Learning
Author Information
Author(s): Troy M. Houser
Primary Institution: University of Oregon
Hypothesis
Can a boundedly rational model effectively capture key aspects of human category learning and memory?
Conclusion
The BR-DIVA model captures the classical order of difficulty observed in category learning and predicts lower accuracy on Type 5 than previously thought.
Supporting Evidence
- BR-DIVA predicts lower accuracy on Type 5 than DIVA.
- BR-DIVA captures the classical order of difficulty observed in the Six Problems.
- Empirical data supports BR-DIVA's predictions better than DIVA's.
Takeaway
This study created a computer model to understand how people learn to categorize things, showing that some categories are harder to learn than others.
Methodology
The study simulated 100 participants performing the Six Problems of category learning using a variational autoencoder model.
Limitations
The study only tested one category learning paradigm with simple stimuli, limiting the generalizability of the findings.
Participant Demographics
Simulated participants based on previous empirical data.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website