SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.02664 · NeurIPS · 2023

Structure-free Graph Condensation: From Large-scale Graphs to Condensed Graph-free Data

Shirui Pan, Xin Zheng, Miao Zhang, Chunyang Chen, Quoc Viet, Hung Nguyen, Xingquan Zhu, 𝜽 𝒕, 𝟎 %𝟏, 𝟎 %𝒒, ℒ 𝐦𝐞𝐭𝐚, 𝐭𝐭, and 3 more

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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 4 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/2306.02664.json · badge: /api/ran/2306.02664.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
Amanda-Zheng/SFGC canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
amanda-zheng/sfgc community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
GNTK Ran
ran on a synthesized input
amanda-zheng/sfgc/gntk_cond.py
code served (permissive licence) · get_code("f1944b965af049e0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 numpy as np
import scipy as sp
import math

INIT_ARGS = {
    'num_layers': 2,
    'num_mlp_layers': 1,
    'jk': False,
    'scale': 'uniform'
}

ARGS = {
    'feat1': np.random.randn(3, 4).astype(np.float64),
    'feat2': np.random.randn(2, 4).astype(np.float64),
    'diag_list1': [np.ones(3), np.ones(3)],
    'diag_list2': [np.ones(2), np.ones(2)],
    'A1': np.array([[0, 1, 1], [1, 0, 1], [1, 1, 0]], dtype=np.float64),
    'A2': np.array([[0, 1], [1, 0]], dtype=np.float64)
}

METHOD = 'gntk'

if __name__ == '__main__':
    from gntk_module import GNTK
    
    gntk = GNTK(**INIT_ARGS)
    
    diag_list1 = gntk.diag(ARGS['feat1'], ARGS['A1'])
    diag_list2 = gntk.diag(ARGS['feat2'], ARGS['A2'])
    
    result = gntk.gntk(
        ARGS['feat1'],
        ARGS['feat2'],
        diag_list1,
        diag_list2,
        ARGS['A1'],
        ARGS['A2']
    )
    
    print("GNTK result shape:", result.shape if hasattr(result, 'shape') else type(result))
    print("GNTK computation completed successfully")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
gntk_module, numpy, scipy — 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
calc Ran
ran on a synthesized input
Amanda-Zheng/SFGC/metagtt_inductive_adj_identity.py
code served (permissive licence) · get_code("0335b850684b8c75")
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 (2, 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
import numpy as np

torch.manual_seed(0)

class MockGNTK(nn.Module):
    def __init__(self):
        super().__init__()
    
    def gntk(self, feat1, feat2, diag1, diag2, A1, A2):
        # Return a small tensor based on input shapes
        batch_size = feat1.shape[0]
        return torch.randn(batch_size, batch_size)

# Create mock GNTK object
gntk = MockGNTK()

# Create small tensors with distinct dimensions
# feat1, feat2: feature tensors (batch_size=2, feature_dim=4)
feat1 = torch.randn(2, 4)
feat2 = torch.randn(2, 4)

# diag1, diag2: diagonal tensors (batch_size=2, dim=3)
diag1 = torch.randn(2, 3)
diag2 = torch.randn(2, 3)

# A1, A2: matrix tensors (batch_size=2, matrix_dim=5, matrix_dim=5)
A1 = torch.randn(2, 5, 5)
A2 = torch.randn(2, 5, 5)

ARGS = {
    "gntk": gntk,
    "feat1": feat1,
    "feat2": feat2,
    "diag1": diag1,
    "diag2": diag2,
    "A1": A1,
    "A2": A2
}
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 Tensor, ndim 2, dtype family float, shape (2, 2) — shape (2, 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
loss_acc_fn_train Ran
ran on a synthesized input
Amanda-Zheng/SFGC/metagtt_inductive_adj_identity.py
code served (permissive licence) · get_code("dace48b54743b457")
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 float; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 numpy as np

torch.manual_seed(0)

# Create a small stand-in data class
class Data:
    def __init__(self):
        self.idx_train = np.array([0, 1])

# Create arguments
data = Data()

# k_ss: support kernel matrix (n_support x n_support)
# Based on function usage: needs to be square, used in trace and inv
n_support = 4
k_ss = torch.randn(n_support, n_support, dtype=torch.float64)

# k_ts: target-support kernel matrix (n_total x n_support)
# Used as k_ts[data.idx_train, :] so needs at least 2 rows
n_total = 5
k_ts = torch.randn(n_total, n_support, dtype=torch.float64)

# y_support: support labels (numpy array)
# Used in matmul with shape (n_support,) or (n_support, 1)
y_support = np.random.randn(n_support).astype(np.float64)

# y_target: target labels (numpy array)
# Used in mse_loss, should match pred shape which is (len(idx_train),)
y_target = np.random.randn(len(data.idx_train)).astype(np.float64)

# reg: regularization parameter
reg = 0.05

ARGS = {
    'data': data,
    'k_ss': k_ss,
    'k_ts': k_ts,
    'y_support': y_support,
    'y_target': y_target,
    'reg': reg
}
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 tuple, ndim 1, 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
one_hot Ran
ran on a synthesized input
Amanda-Zheng/SFGC/metagtt_inductive_adj_identity.py
code served (permissive licence) · get_code("f49468c140bc342f")
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 2, dtype family float, shape (4, 3) 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 one_hot function
x = np.array([0, 2, 1, 0], dtype=np.int64)  # 1D array of class indices
num_classes = 3
center = True
dtype = np.float32

ARGS = {
    "x": x,
    "num_classes": num_classes,
    "center": center,
    "dtype": dtype
}
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 2, dtype family float, shape (4, 3) — shape (4, 3) 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

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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

Graph condensation, which reduces the size of a large-scale graph by synthesizing a small-scale condensed graph as its substitution, has immediate benefits for various graph learning tasks. However, existing graph condensation methods rely on the joint optimization of nodes and structures in the condensed graph, and overlook critical issues in effectiveness and generalization ability. In this paper, we advocate a new Structure-Free Graph Condensation paradigm, named SFGC, to distill a largescale graph into a small-scale graph node set without explicit graph structures, i.e., graph-free data. Our idea is to implicitly encode topology structure information into the node attributes in the synthesized graph-free data, whose topology is reduced to an identity matrix. Specifically, SFGC contains two collaborative components: (1) a training trajectory meta-matching scheme for effectively synthesizing smallscale graph-free data; (2) a graph neural feature score metric for dynamically evaluating the quality of the condensed data. Through training trajectory metamatching, SFGC aligns the long-term GNN learning behaviors between the largescale graph and the condensed small-scale graph-free data, ensuring comprehensive and compact transfer of informative knowledge to the graph-free data. Afterward, the underlying condensed graph-free data would be dynamically evaluated with the graph neural feature score, which is a closed-form metric for ensuring the excellent expressiveness of the condensed graph-free data. Extensive experiments verify the superiority of SFGC across different condensation ratios. ‡ † Corresponding author ‡ Code is available at https://github.com/Amanda-Zheng/SFGC 37th Conference on Neural Information Processing Systems (NeurIPS 2023).

For agents

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

get_harvested_code_for_paper("2306.02664")
get_code_for_paper("2306.02664")
have("2306.02664")

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

curl https://syntology.ai/api/ran/2306.02664.json

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

[![Syntology run record](https://syntology.ai/api/ran/2306.02664.svg)](https://syntology.ai/paper/2306.02664)

Connect an agent — have() is free.