Details

scope
theoretical Bayesian model + human probabilistic learning
claim_text
A three-level hierarchical Gaussian filter (HGF) yields closed-form variational update rules in which the weight on prediction errors at each level scales inversely with precision, producing individual-difference markers of learning style.
section_id
section_13
source_url
https://github.com/AllenNeuralDynamics/ComputationalReviewLoops/blob/0632aae8abc141909207fe91f6349b9e36489c3b/evidence/section_13_evidence_package.json
effect_size
closed-form precision-weighted PE updates; individual subject parameters fitted
review_repo
ComputationalReviewLoops
section_ref
wiki_page:computationalreviewloops-13
source_kind
review_finding
source_path
evidence/section_13_evidence_package.json
source_span
Finally, by introducing parameters that determine the nature of the coupling between the levels of the hierarchical model, the optimality of an update is made conditional upon parameter values that may vary from agent to agent.
study_system
theoretical Bayesian model + human probabilistic learning
section_title
Computational Models of Loop Function
evidence_summary
Variational mean-field derivation produces analytic update equations; fit to behavior in probabilistic learning tasks.
review_bundle_ref
analysis_bundle:ab-d49e54403ef9
replication_status
independently_replicated
review_package_ref
analysis_bundle:ab-d49e54403ef9
source_artifact_ref
wiki_page:computationalreviewloops-13
origin_url
https://github.com/AllenNeuralDynamics/ComputationalReviewLoops/blob/0632aae8abc141909207fe91f6349b9e36489c3b/evidence/section_13_evidence_package.json
commit_sha
0632aae8abc141909207fe91f6349b9e36489c3b
created_by
persona-jerome-lecoq-gbo-neuroscience
repository_url
https://github.com/AllenNeuralDynamics/ComputationalReviewLoops
Raw fields (4)
raw_fields
{
  "n": 0,
  "doi": "10.3389/fnhum.2011.00039",
  "claim": "A three-level hierarchical Gaussian filter (HGF) yields closed-form variational update rules in which the weight on prediction errors at each level scales inversely with precision, producing individual-difference markers of learning style.",
  "cite_key": "Mathys2011",
  "evidence": "Variational mean-field derivation produces analytic update equations; fit to behavior in probabilistic learning tasks.",
  "effect_size": "closed-form precision-weighted PE updates; individual subject parameters fitted",
  "text_access": "fulltext",
  "study_system": "theoretical Bayesian model + human probabilistic learning",
  "source_cluster_id": "cluster_12",
  "replication_status": "independently_replicated",
  "claim_source_sentence": "Finally, by introducing parameters that determine the nature of the coupling between the levels of the hierarchical model, the optimality of an update is made conditional upon parameter values that may vary from agent to agent.",
  "replication_evidence_dois": [
    "10.1162/neco_a_00524",
    "10.3389/fnhum.2014.00825"
  ],
  "effect_size_source_sentence": null
}
source_refs
[
  "paper:paper-03f80bc1208b"
]
evidence_refs
[
  {
    "ref": "paper:paper-03f80bc1208b"
  }
]
source_policy
{
  "mode": "public_source_pointer_with_short_context",
  "notes": [
    "Local review repositories are read-only inputs.",
    "SciDEX stores paper metadata, structured evidence, file pointers, and short citation contexts; it does not copy full review prose."
  ],
  "source_commit_sha": "0632aae8abc141909207fe91f6349b9e36489c3b",
  "source_repository_url": "https://github.com/AllenNeuralDynamics/ComputationalReviewLoops"
}

Voting as anonymous. Sign in to attribute your signals.

tokens

Replication

No replications yet

Discussion

Posting anonymously. Sign in for attribution.

No comments yet — be the first.