SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2606.01340 · 2026

Sample Complexity and Decision-Theoretic Guarantees for Bayesian Model Averaging over Decision Trees with Catalan-Exponential Priors

Livija Jakaite

arXiv · PDF

Code that ran

15 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 4 honoured the contract we drafted, 5 violated it and 0 have no contract verdict; 6 failures of our instrument; 1 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2606.01340.json · badge: /api/ran/2606.01340.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
vitsch/jbdt canonical (classifier label)
classifier basis: census
15 ran · 1 no recorded run
FunctionStatusWhere it lives
bma_weights_from_waic Ran
ran, violated the contract we drafted
vitsch/jbdt/src/bma_zernike.py
pointer only (licence: NONE) · get_code("a5b6e39249dc93f9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

catalan_moments Ran
ran; our contract draft was wrong, not the code
vitsch/jbdt/src/catalan_prior.py
pointer only (licence: NONE) · get_code("352c3e62169fc424")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

catalan_pmf Ran
ran, honoured the contract we drafted
vitsch/jbdt/src/catalan_prior.py
pointer only (licence: NONE) · get_code("ef94a055dc197b28")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

cost_curve Ran
ran, violated the contract we drafted
vitsch/jbdt/src/decision_theory.py
pointer only (licence: NONE) · get_code("f3f9641fcea1ea9c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

dm_posterior_mean Ran
ran, violated the contract we drafted
vitsch/jbdt/src/bma_oracle.py
pointer only (licence: NONE) · get_code("23abfe0243ce1108")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

expected_cost Ran
ran, violated the contract we drafted
vitsch/jbdt/src/decision_theory.py
pointer only (licence: NONE) · get_code("e7296f6d88b1eee0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

feature_slice Ran
ran; our contract draft was wrong, not the code
vitsch/jbdt/benchmark.py
pointer only (licence: NONE) · get_code("b47fb0f8a188d4c4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

feature_slice Ran
ran; our contract draft was wrong, not the code
vitsch/jbdt/benchmark2.py
pointer only (licence: NONE) · get_code("bceae47d14d4a01f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

kl_divergence Ran
ran, violated the contract we drafted
vitsch/jbdt/src/bma_oracle.py
pointer only (licence: NONE) · get_code("3c3e6a0f9dde08a3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

log_catalan_prior_unnorm Ran
ran, honoured the contract we drafted
vitsch/jbdt/src/catalan_prior.py
pointer only (licence: NONE) · get_code("4bc3a286f93ecd2f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

n_90 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vitsch/jbdt/catalan_concentration.py
pointer only (licence: NONE) · get_code("e5243e3e69244a52")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

n_thresh Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vitsch/jbdt/catalan_concentration.py
pointer only (licence: NONE) · get_code("53ee747d638bbd6b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

n_zernike Ran
ran, honoured the contract we drafted
vitsch/jbdt/src/bma_zernike.py
pointer only (licence: NONE) · get_code("a48c8cccc61957e8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

optimal_threshold Ran
ran, honoured the contract we drafted
vitsch/jbdt/src/decision_theory.py
pointer only (licence: NONE) · get_code("f46c01e81389a425")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

oracle_assignment Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vitsch/jbdt/src/bma_oracle.py
pointer only (licence: NONE) · get_code("3ddaf23b775b9e99")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

load_image Not yet run
unverified (harvested, no recorded run)
vitsch/jbdt/benchmark.py
pointer only (licence: NONE) · get_code("5056279c0594a577")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, vitsch/jbdt when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 16 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We ask: when do Bayesian model averaging (BMA) weights over decision trees carry sufficient epistemic information to justify committed exploitation of the averaging distribution? We answer this question in closed form for Bayesian decision trees (BDTs) with Dirichlet-Multinomial leaf models and a Catalan-exponential tree-size prior (Schetinin and Jakaite, 2025), establishing a complete non-asymptotic theory of rational commitment thresholds. Four results are established. First, we prove that WAIC and leave-one-out cross-validation deviances agree up to O(N -1 ) for DM-leaf models, yielding a finite-sample commitment threshold N min ≈ 5.41/∆ (Theorem 3.4) that exactly explains the empirical BMA failure on the N = 40 knee osteoarthritis dataset of Jakaite et al. ( 2021): at N = N min , BMA weights are maximally epistemically fragile. Second, the Catalan-exponential prior's tail decays at rate (e -γ /4) k0 -doubly faster than the Chipman prior-and posterior concentration on the true complexity k * is achieved at N = 300 for Catalan (γ = 1) versus N > 800 for Chipman (Theorem 4.2). Third, WAIC-weighted BMA satisfies an oracle inequality with excess risk O(exp(-N ∆ WAIC /2)) and BMA weight entropy H(w) collapses at the same exponential rate (Theorem 5.1, Proposition 8.1). Fourth, the PAC-Bayes commitment cost N min = -log π(k * )/(2ε 2 ) is 8.1× smaller under the Catalan prior than Chipman for sparse true models (k * = 1, Theorem 6.1): prior design directly governs when rational exploitation begins. These results provide the first tractable, closed-form realisation of the entropy-modulated commitment criterion EU λ (a) = EU(a) -λ(H) • commit(a) (Ortega and Braun, 2013;Grünwald and Dawid, 2004), instantiating its λ-transition boundary in the conjugate-prior case. All results are verified by simulation; three new figures trace the entropy collapse trajectory and commitment-cost surface. All code and precomputed results needed to reproduce the experiments and figures are publicly available at https://github.com/vitsch/jbdt.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2606.01340")
get_code_for_paper("2606.01340")
have("2606.01340")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2606.01340.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2606.01340.svg)](https://syntology.ai/paper/2606.01340)

Connect an agent — have() is free.