메뉴
Distributional Active Inference — push-forward returns, quantile/IQN updates, VMP/Bethe FE, distributional case assignment, policy G(π), ERP profiles, convergence metrics. Use for manuscript §7c and quantitative DAIF results.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Linguistic case formally typed as an analytic category — CaseRole, Morphism, AlignmentFunctor, MonoidalFunctor, Natural Transformations. Crucial for §2 core topology, downstream §7c DAIF alignment, and §9b adversarial injection formal verification.
Scalar active inference on case diagrams — beliefs, variational free energy, belief updates, prediction error, EFE, N400/P600 proxies. Use for §7; use src.daif for distributional returns and full ERP pipeline (§7c).
DisCoCat/DisCoCirc-style string diagrams, complexity metrics, ditransitive constructions, discourse-level prompt scanning. Use for manuscript §3–§4c (and cross-links to §5b magnitude homology metrics, §9b protocol-level checks where `Discourse` is used).
[0,1]-enriched categories, standard proximity matrix, categorical magnitude helpers. Use for §5 distributional proximity and enrichment that feeds cognitive/security modules.
POVM-based quantum case assignment — CasePOVM, case probabilities, crisp/graded/fluid-S POVM factories, semantic state helpers. Use for §8 and §8b quantum semantics.
Cognitive security — type violations, CaseFrameValidator, injection scoring, topological robustness, semantic integrity. Use for §9b adversarial and type-checking narratives.
| name | ccd-daif |
| description | Distributional Active Inference — push-forward returns, quantile/IQN updates, VMP/Bethe FE, distributional case assignment, policy G(π), ERP profiles, convergence metrics. Use for manuscript §7c and quantitative DAIF results. |
src/daif/ERPProfile, policy and diagnostics.from src.daif import (
DistributionalReturn, DAIFResult, ERPProfile,
push_forward_return, distributional_bellman_operator, categorical_return_distribution,
quantile_td_update, implicit_quantile_network_update, wasserstein_return_distance,
distributional_case_assignment, variational_message_passing,
bethe_free_energy, expected_information_gain,
distributional_prediction_error, n400_from_return_distribution,
p600_from_precision_update, erp_amplitude_profile,
G_policy, softmax_policy_selection, distributional_epistemic_value,
convergence_diagnostics, distributional_kl, quantile_coverage,
return_distribution_entropy,
)
§7c (07c_daif_results.md).