SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2502.18237 · ICLR · 2025

Beyond the convexity assumption: Realistic tabular data generation under quantifier-free real linear constraints

Mihaela Stoian, Eleonora Giunchiglia

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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; 0 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2502.18237.json · badge: /api/ran/2502.18237.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
mihaela-stoian/drl_dgm canonical (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
any_disjunctions_in_constraint_set Ran
ran on a synthesized input
mihaela-stoian/drl_dgm/DRL/constraints_code/utils_functions.py
code served (permissive licence) · get_code("06c595f331c18d3c")
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 bool, 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 torch.nn as nn

class Constraint:
    def __init__(self, list_inequalities):
        self.list_inequalities = list_inequalities

# Create test constraints
constraint_no_disjunction = Constraint([])
constraint_single_inequality = Constraint([1])
constraint_with_disjunction = Constraint([1, 2, 3])

# Build the arguments
ARGS = {
    "constraints": [constraint_no_disjunction, constraint_single_inequality, constraint_with_disjunction]
}
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 bool, 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
eval_atoms_list Ran
ran on a synthesized input
mihaela-stoian/drl_dgm/DRL/constraints_code/utils_functions.py
code served (permissive licence) · get_code("424476321619b334")
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 1, dtype family float, shape (2,) 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 torch.nn as nn
from typing import List

torch.manual_seed(0)

# Define a minimal Atom class that the function expects
class Variable:
    def __init__(self, id):
        self.id = id

class Atom:
    def __init__(self, variable_id):
        self.variable = Variable(variable_id)
    
    def eval(self, atom_value):
        # atom_value is a 1D tensor of shape (batch_size,)
        # Return a tensor of the same shape
        return atom_value * 2.0  # Simple evaluation function

# Create test data
batch_size = 2
num_vars = 4

# Create atoms_list with atoms referencing different variables
atoms_list = [
    Atom(variable_id=0),
    Atom(variable_id=1),
    Atom(variable_id=2),
]

# Create preds tensor of shape (batch_size, num_vars)
preds = torch.randn(batch_size, num_vars)

# Define ARGS
ARGS = {
    'atoms_list': atoms_list,
    'preds': preds,
    'reduction': 'sum'
}
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 1, dtype family float, shape (2,) — shape (2,) 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
gen_sat_check Ran
ran on a synthesized input
mihaela-stoian/drl_dgm/DRL/evaluation/eval_for_testing.py
code served (permissive licence) · get_code("32a45193161c760a")
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. 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 torch
import pandas as pd
import numpy as np

torch.manual_seed(0)

# Create a mock constraint class with disjunctive_inequality
class MockDisjunctiveInequality:
    def check_satisfaction(self, data):
        # Returns a boolean tensor indicating satisfaction per datapoint
        # For this mock, we'll say all datapoints satisfy this constraint
        return torch.ones(data.shape[0], dtype=torch.bool)

class MockConstraint:
    def __init__(self):
        self.disjunctive_inequality = MockDisjunctiveInequality()

# Create args dictionary
ARGS = {
    'args': type('Args', (), {'debug': False})(),  # Simple args object
    'generated_data': {
        'train': [
            pd.DataFrame(np.random.randn(2, 4), columns=['a', 'b', 'c', 'd']),
            pd.DataFrame(np.random.randn(2, 4), columns=['a', 'b', 'c', 'd'])
        ]
    },
    'constraints': [
        MockConstraint(),
        MockConstraint(),
        MockConstraint()
    ],
    'log_wandb': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, torch, wandb — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
get_constr_at_level_x Ran
ran on a synthesized input
mihaela-stoian/drl_dgm/DRL/constraints_code/correct_predictions.py
code served (permissive licence) · get_code("4755d8451438ac43")
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

class ConstraintVar:
    def __init__(self, id_val):
        self.id = id_val

torch.manual_seed(0)

# Create constraint variable objects with unique ids
var1 = ConstraintVar(id_val=1)
var2 = ConstraintVar(id_val=2)
var3 = ConstraintVar(id_val=3)

# Create the sets_of_constr dictionary mapping variables to their constraints
sets_of_constr = {
    var1: ["constraint_a", "constraint_b"],
    var2: ["constraint_c"],
    var3: ["constraint_d", "constraint_e", "constraint_f"]
}

# Create the x variable we're looking for
x = var2

ARGS = {
    "x": x,
    "sets_of_constr": sets_of_constr
}
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 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
real_sat_check Ran
ran on a synthesized input
mihaela-stoian/drl_dgm/DRL/evaluation/eval_for_testing.py
code served (permissive licence) · get_code("fefb0090b04fdf60")
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. 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 torch
import pandas as pd

torch.manual_seed(0)

class DisjunctiveInequality:
    def check_satisfaction(self, data):
        # Returns a boolean tensor indicating satisfaction for each sample
        return torch.ones(data.shape[0], dtype=torch.bool)

class Constraint:
    def __init__(self):
        self.disjunctive_inequality = DisjunctiveInequality()

# Create small real data
real_data_numpy = pd.DataFrame({
    'col1': [1.0, 2.0],
    'col2': [3.0, 4.0],
    'col3': [5.0, 6.0],
    'col4': [7.0, 8.0]
})

real_data = {'train': real_data_numpy}

# Create constraints
constraints = [
    Constraint(),
    Constraint()
]

ARGS = {
    'args': None,
    'real_data': real_data,
    'constraints': constraints,
    'log_wandb': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, torch, wandb — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 5 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

Synthetic tabular data generation has traditionally been a challenging problem due to the high complexity of the underlying distributions that characterise this type of data. Despite recent advances in deep generative models (DGMs), existing methods often fail to produce realistic datapoints that are well-aligned with available background knowledge. In this paper, we address this limitation by introducing Disjunctive Refinement Layer (DRL), a novel layer designed to enforce the alignment of generated data with the background knowledge specified in user-defined constraints. DRL is the first method able to automatically make deep learning models inherently compliant with constraints as expressive as quantifier-free linear formulas, which can define non-convex and even disconnected spaces. Our experimental analysis shows that DRL not only guarantees constraint satisfaction but also improves efficacy in downstream tasks. Notably, when applied to DGMs that frequently violate constraints, DRL eliminates violations entirely. Further, it improves performance metrics by up to 21.4% in F1-score and 20.9% in Area Under the ROC Curve, thus demonstrating its practical impact on data generation.

For agents

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

get_harvested_code_for_paper("2502.18237")
get_code_for_paper("2502.18237")
have("2502.18237")

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

curl https://syntology.ai/api/ran/2502.18237.json

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

[![Syntology run record](https://syntology.ai/api/ran/2502.18237.svg)](https://syntology.ai/paper/2502.18237)

Connect an agent — have() is free.