Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Learning overhypotheses with hierarchical Bayesian models.

Charles Kemp1, Amy Perfors, Joshua B Tenenbaum

  • 1Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA. ckemp@mit.edu

Developmental Science
|April 21, 2007
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Conniving With Continuations: Representing Goals in a Domain-Specific Language of Thought.

Topics in cognitive science·2026
Same author

Neural representation of action symbols in primate frontal cortex.

Nature·2026
Same author

Human-level learning of complex novel tasks as theory-based modelling, exploration and planning.

Philosophical transactions. Series A, Mathematical, physical, and engineering sciences·2026
Same author

Reverse engineering the centered self.

Psychological review·2026
Same author

Symmetry in category systems across languages.

Nature communications·2026
Same author

What people learn from punishment: A cognitive model.

Proceedings of the National Academy of Sciences of the United States of America·2025
Same journal

Effects of Children's Video Gaming Experience on Chinese Orthographic Processing and Its Neural Mechanisms.

Developmental science·2026
Same journal

The Kia Tīmata Pai Randomized Controlled Trial: ENRICH Early Childhood Teacher Training Improves Toddlers' Oral Language and Self-Regulation.

Developmental science·2026
Same journal

Deaf Children's Use of Mutual Exclusivity and Eye Gaze to Determine Word Meanings in American Sign Language.

Developmental science·2026
Same journal

Adolescents and Adults Use Different Facial Movements and Vocal Cues to Express Emotions.

Developmental science·2026
Same journal

Decoding Preschool Social Dynamics: Automated Tracking of Spatial and Temporal Patterns to Investigate Social Interactions and Relationships in Peer Groups.

Developmental science·2026
Same journal

Early Life Stress Affects Human Decision Making by Increasing Expectations of Volatility.

Developmental science·2026
See all related articles

Inductive learning requires overhypotheses, or constraints. Hierarchical Bayesian models explain how learners acquire these constraints, including those about feature variability and category grouping.

Area of Science:

  • Cognitive Science
  • Machine Learning
  • Developmental Psychology

Background:

  • Inductive learning, the process of generalizing from specific examples, is fundamental to human and artificial intelligence.
  • Overhypotheses, or prior constraints on possible generalizations, are necessary for successful inductive learning.
  • Existing theories acknowledge the need for innate overhypotheses, but the acquisition of learned overhypotheses remains less understood.

Purpose of the Study:

  • To propose a computational framework for acquiring learned overhypotheses in inductive learning.
  • To demonstrate how hierarchical Bayesian models can explain the acquisition of specific types of overhypotheses.
  • To provide a unified account for both innate and acquired overhypotheses in learning.

Main Methods:

Related Experiment Videos

  • Development of hierarchical Bayesian models to simulate learning processes.
  • Modeling the acquisition of overhypotheses related to feature variability (e.g., shape bias).
  • Modeling the acquisition of overhypotheses concerning the ontological grouping of categories (e.g., objects vs. substances).

    • Main Results:

    • The proposed models successfully acquired relevant overhypotheses from data, mirroring human learning patterns.
    • Demonstrated that hierarchical Bayesian inference can explain the emergence of biases like the shape bias in word learning.
    • Showcased the ability of the models to learn category groupings consistent with ontological distinctions.

    Conclusions:

    • Hierarchical Bayesian models offer a powerful mechanism for explaining the acquisition of learned overhypotheses.
    • This framework helps bridge the gap between innate biases and learned constraints in inductive learning.
    • The findings have implications for understanding cognitive development, artificial intelligence, and the nature of knowledge acquisition.