SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2607.09893 · 2026

An End-to-End Hybrid Quantum-Classical Sampling Workflow for Discrete Markov Random Fields: A Reproducible Case Study

Arul Rhik Mazumder

arXiv · PDF · Open in the Atlas

Code that ran

13 of 14 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; 10 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 1 failure 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/2607.09893.json · badge: /api/ran/2607.09893.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
arulrhikm/QuantumDGM canonical (classifier label)
classifier basis: census
13 ran · 1 no recorded run
FunctionStatusWhere it lives
chain_graph Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/graph_families.py
code served (permissive licence) · get_code("3ac206d9f73fb938")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Graph. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import networkx as nx

def chain_graph(n: int):
    return nx.path_graph(n)

# Construct valid arguments
ARGS = {
    "n": 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Graph — a Graph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Graph output the recorded ndim 1 and dtype family int describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
energy_from_pairwise_mrf Ran
ran; our contract draft was wrong, not the code
arulrhikm/QuantumDGM/src/amplitude_encoding.py
code served (permissive licence) · get_code("88bf488146fe13ca")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - n: int, number of nodes
# - mrf_params: dict with keys that are tuples of node indices and values that are theta parameters
# - config_bits: list[int] of length n, representing the configuration of bits for each node

n = 5

# Create MRF parameters dict
# Keys are tuples of node indices (representing pairwise connections)
# Values are theta parameters (weights)
mrf_params = {
    (0, 1): 0.5,
    (1, 2): -0.3,
    (2, 3): 0.7,
    (0, 3): -0.2,
    (1, 3): 0.4,
}

# Create config_bits - a list of binary values (0 or 1) for each node
config_bits = [1, 0, 1, 0, 1]

ARGS = {
    "n": n,
    "mrf_params": mrf_params,
    "config_bits": config_bits,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
erdos_renyi Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/graph_families.py
code served (permissive licence) · get_code("e0626d736e62fb29")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Graph. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import networkx as nx

ARGS = {
    "n": 10,
    "p": 0.3,
    "seed": 0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Graph — a Graph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Graph output the recorded ndim 1 and dtype family int describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
fidelity Ran
ran, violated the contract we drafted
arulrhikm/QuantumDGM/src/metrics.py
code served (permissive licence) · get_code("30ca5eb2f4a10377")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid probability distributions for the fidelity function
# The function computes quantum fidelity between two probability distributions
# p and q should be 1D arrays representing probability distributions

# Create two simple probability distributions
p = np.array([0.2, 0.3, 0.3, 0.2], dtype=np.float64)
q = np.array([0.25, 0.25, 0.25, 0.25], dtype=np.float64)

ARGS = {
    "p": p,
    "q": q,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_chi_matching_vqc_params Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/bq_mps_baseline.py
code served (permissive licence) · get_code("8f950c9c7950b582")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# find_chi_matching_vqc_params takes:
# - n_vqc_params: int (number of VQC parameters to match)
# - n: int (number of qubits/sites)
# - chi_cap: int (optional, default 512, maximum chi value to search)
#
# The function searches for the smallest chi value where mps_param_count(n, chi) >= n_vqc_params
# It loops through chi from 1 to chi_cap and checks the condition

# Create reasonable test values:
# Let's use a case where we can find a chi value within reasonable range
n_vqc_params = 1000  # target number of VQC parameters
n = 5  # number of sites/qubits
chi_cap = 512  # maximum chi to search

ARGS = {
    "n_vqc_params": n_vqc_params,
    "n": n,
    "chi_cap": chi_cap,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_entanglement_pairs Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/src/entanglement.py
code served (permissive licence) · get_code("ec89bcdafa7bb84d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 2, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import networkx as nx

torch.manual_seed(0)

# Test with 'linear' strategy
ARGS = {
    'strategy': 'linear',
    'n': 5,
    'graph': None
}

# Alternative configurations for testing other strategies:
# For 'full' strategy:
# ARGS = {
#     'strategy': 'full',
#     'n': 4,
#     'graph': None
# }

# For 'clique' strategy with a graph:
# graph = nx.Graph()
# graph.add_edges_from([(0, 1), (1, 2), (2, 0)])
# ARGS = {
#     'strategy': 'clique',
#     'n': 3,
#     'graph': graph
# }
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
grid_graph Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/graph_families.py
code served (permissive licence) · get_code("9a36f5b174ec21db")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Graph. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import networkx as nx

ARGS = {
    'n': 4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Graph — a Graph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Graph output the recorded ndim 2 and dtype family int describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
kl_divergence Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/src/metrics.py
code served (permissive licence) · get_code("b0cf13770f3cffcd")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create two probability distributions
p = np.array([0.1, 0.2, 0.3, 0.4])
q = np.array([0.15, 0.25, 0.35, 0.25])

# Define epsilon parameter
epsilon = 1e-12

# Construct the arguments dictionary
ARGS = {
    "p": p,
    "q": q,
    "epsilon": epsilon
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mean_std Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/plotting.py
code served (permissive licence) · get_code("d63a0e4a54beeeeb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid arguments for mean_std function
# The function takes xs and ys, converts them to arrays, finds unique values in xs,
# and computes mean and std of ys grouped by xs values.

xs = np.array([1, 1, 2, 2, 3, 3])
ys = np.array([10.0, 12.0, 20.0, 22.0, 30.0, 32.0])

ARGS = {
    'xs': xs,
    'ys': ys
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mps_param_count Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/experiments/bq_mps_baseline.py
code served (permissive licence) · get_code("96dd4d96f94fe0e3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# def mps_param_count(n: int, chi_max: int) -> int:
#     return int(2 * chi_max ** 2 * max(1, n - 2) + 4 * chi_max)
#
# Parameters:
# - n: int (appears to be a size/count parameter)
# - chi_max: int (appears to be a maximum chi/bond dimension)
#
# The function performs arithmetic operations only, so we just need valid integers.

ARGS = {
    "n": 5,
    "chi_max": 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
pairwise_probs_from_params Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/src/amplitude_encoding.py
code served (permissive licence) · get_code("59c0fb605c84645b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (8,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct arguments for pairwise_probs_from_params
n = 3
mrf_params = {
    (0, 1): 0.5,
    (1, 2): -0.3,
    (0,): 0.2,
    (2,): -0.1
}

ARGS = {
    "n": n,
    "mrf_params": mrf_params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (8,) — shape (8,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
random_init Ran
ran, honoured the contract we drafted
arulrhikm/QuantumDGM/src/entanglement.py
code served (permissive licence) · get_code("75a98560be497f8e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (15,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# The function random_init takes three parameters:
# - n: int - number of elements
# - depth: int - depth parameter
# - seed: int - random seed (default 0)
# The function uses these to create a uniform random array of size n*depth

ARGS = {
    "n": 5,
    "depth": 3,
    "seed": 42
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (15,) — shape (15,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
tv_distance Ran
ran, violated the contract we drafted
arulrhikm/QuantumDGM/src/metrics.py
code served (permissive licence) · get_code("bd2a97c59532f867")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create two small probability distributions
p = np.array([0.2, 0.3, 0.5], dtype=np.float64)
q = np.array([0.1, 0.4, 0.5], dtype=np.float64)

ARGS = {
    "p": p,
    "q": q,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
savefig Not yet run
unverified (harvested, no recorded run)
arulrhikm/QuantumDGM/experiments/plotting.py
code served (permissive licence) · get_code("35a9fc29c138f312")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 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 14 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

Sampling from discrete Markov random fields (MRFs) is a well-known hard problem in probabilistic inference. We study amplitude-encoded i.i.d. sampling for small discrete MRFs, scoped to the regime where the 2 n target probabilities can be precomputed classically-so no quantum exponential speedup is possible, but the structural property that each circuit execution returns an independent sample (τ ≈ 1) can be cleanly compared against classical MCMC alternatives. Across 60 instances spanning five graph families (barbell, barbell-path, chain, Erdős-Rényi, two-clique) with 1,000-step burn-in and 3,000 retained samples, Quantum/Single-Site-Gibbs, Quantum/Block-Gibbs, Quantum/Tuned-Block-Gibbs, and Quantum/Parallel-Tempering ESS ratios have means of 16.35, 7.29, 1.82, and 1.79 respectively, showing that modern classical samplers substantially close-and may eliminate-the ESS gap relative to amplitude-encoded sampling. When the O(2 n ) classical preprocessing required by the amplitude encoder is amortized into wall-clock time, exact inverse-CDF sampling reaches a mean of 17,683,598 ESS/s versus 487,706 ESS/s for the quantum sampler (36× on mean rates, 153× mean perinstance), confirming no wall-clock advantage in this regime. The contribution is therefore not a speedup claim but (i) a clean characterization of MCMC autocorrelation costs under a fixed protocol and (ii) a reproducible benchmark of amplitudeencoded state preparation for discrete MRFs at n = 8, 10, 12. We further report a multi-trial matrix product state (MPS) scaling study (three seeds per point, n up to 40) showing χ = 32 achieves F = 0.721 ± 0.059 at n = 40, and a matched-budget variational quantum circuit (VQC) vs. MPS comparison at n = 8, 10, 12 where VQC fidelities fall below MPS at every point ((FVQC, FMPS) = (0.306, 0.990), (0.210, 0.958), (0.165, 0.878) at compressions 10.7×, 34.1×, 113.8×)-a negative result for shallow hardware-efficient ansätze. Code and data: https://github. com/arulrhikm/QuantumDGM.

For agents

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

get_harvested_code_for_paper("2607.09893")
get_code_for_paper("2607.09893")
have("2607.09893")

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

curl https://syntology.ai/api/ran/2607.09893.json

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

[![Syntology run record](https://syntology.ai/api/ran/2607.09893.svg)](https://syntology.ai/paper/2607.09893)

Connect an agent — have() is free.