SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.10643 · 2026

Evaluation metrics for temporal preservation in synthetic longitudinal patient data

Antti Airola, Katariina Perkonoja, Parisa Movahedi, Kari Auranen, Joni Virta

arXiv · PDF · Open in the Atlas

Code that ran

4 of 6 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 2 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/2602.10643.json · badge: /api/ran/2602.10643.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
Nic5472K/ScientificData2021_HealthGym canonical (classifier label)
classifier basis: census
3 ran · 1 no recorded run
btheodorou99/HALO_Inpatient community
Papers with Code archive
1 ran · 1 no recorded run
FunctionStatusWhere it lives
Execute_C008 Ran
ran; our contract draft was wrong, not the code
Nic5472K/ScientificData2021_HealthGym/Academic/mimic_III_sepsis/C008_Utils_ReplaceNames.py
code served (permissive licence) · get_code("286559a3696fa045")
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 1, 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 torch
import pandas as pd

torch.manual_seed(0)

# Create a minimal data_types dict that has a 'name' key
# The function accesses list(data_types['name']) and expects it to have at least 45 elements
data_types = {
    'name': [
        'col0', 'col1', 'col2', 'col3', 'col4', 'col5', 'col6', 'col7', 'col8', 'col9',
        'col10', 'col11', 'col12', 'col13', 'col14', 'col15', 'col16', 'col17', 'col18', 'col19',
        'col20', 'col21', 'col22', 'col23', 'col24', 'col25', 'col26', 'col27', 'col28', 'col29',
        'col30', 'col31', 'col32', 'col33', 'col34', 'col35', 'col36', 'col37', 'col38', 'col39',
        'col40', 'col41', 'col42', 'col43', 'col44'
    ]
}

ARGS = {
    'data_types': data_types
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — 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
LoadPreTrain Ran
ran, honoured the contract we drafted
Nic5472K/ScientificData2021_HealthGym/Academic/mimic_III_sepsis/D004_WGAN_GP.py
code served (permissive licence) · get_code("65266aea731d2cf2")
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 1, 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 tempfile
import os

torch.manual_seed(0)

# Create temporary directory for mock state dicts
temp_dir = tempfile.mkdtemp()

# Create mock state dicts and save them
mock_G_state_dict = {
    'layer1.weight': torch.randn(4, 3),
    'layer1.bias': torch.randn(4)
}

mock_D_state_dict = {
    'layer1.weight': torch.randn(3, 2),
    'layer1.bias': torch.randn(3)
}

G_path = os.path.join(temp_dir, 'G_SD.pt')
D_path = os.path.join(temp_dir, 'D_SD.pt')

torch.save(mock_G_state_dict, G_path)
torch.save(mock_D_state_dict, D_path)

# Case 1: Continue = False (returns (0, 0))
ARGS = {
    'content': [False, G_path, D_path, 0]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family int — 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
correlation Ran
ran; our contract draft was wrong, not the code
Nic5472K/ScientificData2021_HealthGym/Academic/mimic_III_sepsis/D004_WGAN_GP.py
code served (permissive licence) · get_code("e95edf3eda6485ff")
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 Tensor, ndim 2, dtype family float, shape (4, 4) 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

torch.manual_seed(0)

# Create a small tensor with shape (2, 4) - batch size 2, feature dimension 4
x = torch.randn(2, 4)

ARGS = {
    "x": x,
    "eps": 1e-08
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (4, 4) — shape (4, 4) 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
generate_statistics Ran
ran; our contract draft was wrong, not the code
btheodorou99/HALO_Inpatient/continuous_variables/evaluate.py
code served (permissive licence) · get_code("08a14d849249ccaa")
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) and returned dict. 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 numpy as np
from tqdm import tqdm

class HALOConfig(object):
    def __init__(self, total_vocab_size=14487, code_vocab_size=14167, lab_vocab_size=237, continuous_vocab_size=15, label_vocab_size=65, special_vocab_size=3, categorical_lab_vocab_size=47, continuous_lab_vocab_size=190, phenotype_labels=25, n_positions=150, n_ctx=150, n_embd=1440, n_layer=12, n_head=18, layer_norm_epsilon=1e-05, initializer_range=0.02, batch_size=56, sample_batch_size=128, epoch=50, lr=0.0001):
        self.total_vocab_size = total_vocab_size
        self.code_vocab_size = code_vocab_size
        self.label_vocab_size = label_vocab_size
        self.lab_vocab_size = lab_vocab_size
        self.categorical_lab_vocab_size = categorical_lab_vocab_size
        self.continuous_lab_vocab_size = continuous_lab_vocab_size
        self.continuous_vocab_size = continuous_vocab_size
        self.special_vocab_size = special_vocab_size
        self.phenotype_labels = phenotype_labels
        self.n_positions = n_positions
        self.n_ctx = n_ctx
        self.n_embd = n_embd
        self.n_layer = n_layer
        self.n_head = n_head
        self.layer_norm_epsilon = layer_norm_epsilon
        self.initializer_range = initializer_range
        self.batch_size = batch_size
        self.sample_batch_size = sample_batch_size
        self.epoch = epoch
        self.lr = lr

config = HALOConfig(label_vocab_size=3)

ehr_dataset = [
    {
        'labels': [0, 1, 0],
        'visits': [
            ([1, 2, 3], [], None, [5.0]),
            ([4, 5], [], None, [10.0]),
        ]
    },
    {
        'labels': [1, 0, 1],
        'visits': [
            ([2, 3, 6], [], None, [3.0]),
            ([7, 8, 9], [], None, [7.0]),
        ]
    },
    {
        'labels': [0, 0, 1],
        'visits': [
            ([1, 4], [], None, [2.0]),
        ]
    },
]

ARGS = {
    'ehr_dataset': ehr_dataset
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float 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
Execute_C003 Not yet run
unverified (harvested, no recorded run)
Nic5472K/ScientificData2021_HealthGym/Academic/mimic_III_sepsis/C003_Utils_Loders.py
code served (permissive licence) · get_code("beaab1e051e25dd3")
How it ran

no invocation recorded.

HALOConfig Not yet run
unverified (harvested, no recorded run)
btheodorou99/HALO_Inpatient/continuous_variables/evaluate.py
code served (permissive licence) · get_code("50bdf4b5fe6460a4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

Novel metrics are proposed to assess temporal preservation in synthetic longitudinal data • The metrics cover mean, covariance, subject-level, and measurement structures • Multi-metric evaluation gives clearer insight into synthetic data quality • Assessing only mean-level resemblance can mask loss of temporal dependencies • Data quality, balance, and preprocessing strongly affect temporal preservation

For agents

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

get_harvested_code_for_paper("2602.10643")
get_code_for_paper("2602.10643")
have("2602.10643")

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

curl https://syntology.ai/api/ran/2602.10643.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.10643.svg)](https://syntology.ai/paper/2602.10643)

Connect an agent — have() is free.