SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1806.08804 · 2018

Hierarchical Graph Representation Learning with Differentiable Pooling

arXiv · PDF · Open in the Atlas

Code that ran

10 of 20 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 9 have no contract verdict; 1 failure of our instrument; 10 no recorded run), measured 2026-11-21; 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/1806.08804.json · badge: /api/ran/1806.08804.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
basiralab/reproduciblefedgnn community (classifier label)
classifier basis not recorded on the link
7 ran · 4 no recorded run
RexYing/diffpool community (classifier label)
classifier basis not recorded on the link
2 ran · 5 no recorded run
chappers/graph-differential-pooling community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
PasqualeAuriemma/GCN-DIFFPOOL reimplementation (classifier label)
classifier basis: classifier_2025
0 ran · 1 no recorded run
FunctionStatusWhere it lives
Mean_W_Cv Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/Analysis.py
code served (permissive licence) · get_code("d13d56b516dc4831")
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 torch
import torch.nn as nn
import numpy as np
import pickle
import tempfile
import os

torch.manual_seed(0)

# Create a temporary directory for mock weight files
temp_dir = tempfile.mkdtemp()

# Create mock pickle files for 5 folds
for fold in range(5):
    # Create directories
    for model_name in ['diffpool', 'sag', 'gunet', 'gat']:
        model_dir = os.path.join(temp_dir, model_name, 'weights')
        os.makedirs(model_dir, exist_ok=True)

# Generate mock weight data and save to pickle files
dataset = 'test_dataset'
view = 0

# For diffpool, gunet, sag models
for i in range(5):
    for model in ['diffpool', 'gunet']:
        filename = os.path.join(
            temp_dir,
            model,
            'weights',
            f'W_MainModel_5Fold_{dataset}_{model}_CV_{i}_view_{view}.pickle'
        )
        # Create mock weight tensor with shape (8, 4) to match axis=1 mean
        mock_dict = {'w': torch.randn(8, 4)}
        with open(filename, 'wb') as f:
            pickle.dump(mock_dict, f)

# For sag model (different filename format)
for i in range(5):
    filename = os.path.join(
        temp_dir,
        'sag',
        'weights',
        f'W_MainModel_5Fold_{dataset}_sag_view_{view}_CV_{i}.pickle'
    )
    mock_dict = {'w': torch.randn(8, 4)}
    with open(filename, 'wb') as f:
        pickle.dump(mock_dict, f)

# For gat model (different structure)
for i in range(5):
    filename = os.path.join(
        temp_dir,
        'gat',
        'weights',
        f'W_MainModel_5Fold_{dataset}_gat_CV_{i}_view_{view}.pickle'
    )
    # For gat/gcn, it accesses [0] on the result, so shape is (1, 4)
    mock_dict = {'w': torch.randn(1, 4)}
    with open(filename, 'wb') as f:
        pickle.dump(mock_dict, f)

# Change to temp directory so relative paths work
original_cwd = os.getcwd()
os.chdir(temp_dir)

# Create a simple stand-in dataset class
class DummyDataset:
    def __init__(self):
        self.name = 'test_dataset'

# Create a simple stand-in model class
class DummyModel:
    def __init__(self):
        pass

dataset_obj = DummyDataset()
model_obj = 'diffpool'  # Pass model as string since function expects string
view_obj = 0

ARGS = {
    'dataset': dataset_obj.name,
    'model': model_obj,
    'view': view_obj
}
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 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
Models_trained Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/Analysis.py
code served (permissive licence) · get_code("0540a7e365607e63")
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 os
import tempfile

# Create temporary directories and files to satisfy the function's os.path.exists checks
temp_dir = tempfile.mkdtemp()

# Create the necessary directory structure and files
models = ['diffpool', 'gat', 'gcn', 'sag', 'gunet']
dataset = 'test_dataset'
view = 1

for model in models:
    model_dir = os.path.join(temp_dir, model, 'weights')
    os.makedirs(model_dir, exist_ok=True)
    
    if model == 'sag':
        cv_filename = f'W_MainModel_5Fold_{dataset}_{model}_view_{view}_CV_4.pickle'
        few_filename = f'W_Few_Shot_{dataset}_{model}_view_{view}_99.pickle'
    else:
        cv_filename = f'W_MainModel_5Fold_{dataset}_{model}_CV_4_view_{view}.pickle'
        few_filename = f'W_Few_Shot_{dataset}_{model}99_view_{view}.pickle'
    
    # Create the files
    open(os.path.join(model_dir, cv_filename), 'w').close()
    open(os.path.join(model_dir, few_filename), 'w').close()

# Change to temp directory so relative paths work
original_dir = os.getcwd()
os.chdir(temp_dir)

ARGS = {
    'dataset': 'test_dataset',
    'view': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
average_weights Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/fed_localmodel.py
code served (permissive licence) · get_code("63d220d8b4f8afd5")
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 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 torch
import copy

torch.manual_seed(0)

# Create a list of weight dictionaries simulating multiple hospitals
# Each hospital has model weights (state_dict-like)
weights = []

# Hospital 1 weights
hospital1_weights = {
    'layer1.weight': torch.randn(4, 3),
    'layer1.bias': torch.randn(4),
    'layer2.weight': torch.randn(5, 4),
    'layer2.bias': torch.randn(5),
}
weights.append(hospital1_weights)

# Hospital 2 weights (same structure, different values)
hospital2_weights = {
    'layer1.weight': torch.randn(4, 3),
    'layer1.bias': torch.randn(4),
    'layer2.weight': torch.randn(5, 4),
    'layer2.bias': torch.randn(5),
}
weights.append(hospital2_weights)

# Hospital 3 weights (same structure, different values)
hospital3_weights = {
    'layer1.weight': torch.randn(4, 3),
    'layer1.bias': torch.randn(4),
    'layer2.weight': torch.randn(5, 4),
    'layer2.bias': torch.randn(5),
}
weights.append(hospital3_weights)

ARGS = {
    'weights': weights,
}
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 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
datasets_splits Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/cross_val.py
code served (permissive licence) · get_code("1f11f715f1139ba1")
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 numpy as np

torch.manual_seed(0)

# Construct folds - a list of lists representing data splits
# Each fold contains sample indices or data
folds = [
    [1, 2, 3],           # fold 0
    [4, 5, 6],           # fold 1
    [7, 8, 9],           # fold 2
    [10, 11, 12]         # fold 3
]

# args parameter - the function doesn't actually use it, but we need to provide it
# It's passed to the function but never accessed in the body
args = {
    "learning_rate": 0.001,
    "batch_size": 32,
    "epochs": 10
}

# val_idx - validation index, selects which fold becomes test set
val_idx = 0

ARGS = {
    "folds": folds,
    "args": args,
    "val_idx": val_idx
}
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 — 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
dense_diff_pool Ran
ran on a synthesized input
chappers/graph-differential-pooling/diff_pool.py
code served (permissive licence) · get_code("2cd5bd6f6bc85e88")
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

torch.manual_seed(0)

# Construct arguments for dense_diff_pool
# Parameters:
# - x: Node feature tensor (B, N, F) - batch_size=2, num_nodes=4, features=3
# - adj: Adjacency tensor (B, N, N) - batch_size=2, num_nodes=4
# - s: Assignment tensor (B, N, C) - batch_size=2, num_nodes=4, clusters=2
# - mask: Optional boolean mask (B, N)

batch_size = 2
num_nodes = 4
feature_dim = 3
num_clusters = 2

# Node features: (batch_size=2, num_nodes=4, feature_dim=3)
x = torch.randn(batch_size, num_nodes, feature_dim)

# Adjacency matrix: (batch_size=2, num_nodes=4, num_nodes=4)
adj = torch.rand(batch_size, num_nodes, num_nodes)
adj = (adj + adj.transpose(1, 2)) / 2  # Make symmetric

# Assignment matrix: (batch_size=2, num_nodes=4, num_clusters=2)
s = torch.randn(batch_size, num_nodes, num_clusters)

# Mask: (batch_size=2, num_nodes=4) - all True to indicate valid nodes
mask = torch.ones(batch_size, num_nodes, dtype=torch.bool)

ARGS = {
    "x": x,
    "adj": adj,
    "s": s,
    "mask": mask,
}
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 — 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
exp_moving_avg Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
RexYing/diffpool/util.py
code served (permissive licence) · get_code("dd691d00aa5cc146")
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 numpy as np

torch.manual_seed(0)

# The function takes x (a sequence) and decay (a float).
# x is indexed with x[0] and iterated with x[1:], so it needs to be a sequence.
# Each element v in x is used in arithmetic operations (shadow - v), 
# so elements should be numeric (int, float, or tensor).

x = [5.0, 4.0, 6.0, 3.0]
decay = 0.9

ARGS = {
    "x": x,
    "decay": decay
}
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 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
gen_2community_ba Ran
ran on a synthesized input
RexYing/diffpool/gen/data.py
code served (permissive licence) · get_code("712b1b3a42060fb0")
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. 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 networkx as nx

class ConstFeatureGen:
    def __init__(self, value):
        self.value = value
    
    def gen_node_features(self, G):
        for node in G.nodes():
            G.nodes[node]['feature'] = self.value

class GaussianFeatureGen:
    def __init__(self, mu, sigma):
        self.mu = mu
        self.sigma = sigma
    
    def gen_node_features(self, G):
        for node in G.nodes():
            G.nodes[node]['feature'] = np.random.multivariate_normal(self.mu, self.sigma)

np.random.seed(0)

n_range = [5, 6, 7]
m_range = [2, 3]
num_graphs = 2
inter_prob = 0.3

mu0 = np.zeros(10)
mu1 = np.ones(10)
sigma0 = np.eye(10) * 0.1
sigma1 = np.eye(10) * 0.1

fg0 = GaussianFeatureGen(mu0, sigma0)
fg1 = GaussianFeatureGen(mu1, sigma1)
feature_generators = [fg0, fg1]

ARGS = {
    'n_range': n_range,
    'm_range': m_range,
    'num_graphs': num_graphs,
    'inter_prob': inter_prob,
    'feature_generators': feature_generators
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list 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_adjs Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/dataLoader_medmnist.py
code served (permissive licence) · get_code("07ed65681aebf2db")
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.

# Setup code for get_adjs function
# This function finds adjacent pixels in an n×n grid

# Test case 1: pixel in the middle of a grid
i_middle = 1
j_middle = 1
n_size = 3

ARGS = {
    "i": i_middle,
    "j": j_middle,
    "n": n_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
img_to_adj Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/dataLoader_medmnist.py
code served (permissive licence) · get_code("3a7c82cbbc34ab03")
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 (16, 16) 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 a small 4x4 square image with integer pixel values
image = np.array([
    [10, 20, 30, 40],
    [50, 60, 70, 80],
    [90, 100, 110, 120],
    [130, 140, 150, 160]
], dtype=np.float32)

ARGS = {
    "image": image
}
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 (16, 16) — shape (16, 16) 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
minmax_sc Ran
ran on a synthesized input
basiralab/reproduciblefedgnn/federated_reproducibility/main_diffpool.py
code served (permissive licence) · get_code("81b03223cd6f4d43")
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 (2, 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 numpy as np
from sklearn import preprocessing

# The function minmax_sc takes a single parameter 'x'
# It expects x to be array-like that can be passed to fit_transform()
# fit_transform() expects a 2D array-like object (samples x features)

# Create a small 2D numpy array (2 samples, 4 features)
x_input = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float32)

ARGS = {
    'x': x_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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 (2, 4) — shape (2, 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
Mean_W_Two_shot Not yet run
unverified (harvested, no recorded run)
basiralab/reproduciblefedgnn/federated_reproducibility/Analysis.py
code served (permissive licence) · get_code("a2ab7f58736f25f8")
How it ran

no invocation recorded.

evaluate Not yet run
unverified (harvested, no recorded run)
basiralab/reproduciblefedgnn/federated_reproducibility/main_diffpool.py
code served (permissive licence) · get_code("1e9d378dbb12714b")
How it ran

no invocation recorded.

evaluate Not yet run
unverified (harvested, no recorded run)
basiralab/reproduciblefedgnn/federated_reproducibility/main_gcn.py
code served (permissive licence) · get_code("96903513f7177254")
How it ran

no invocation recorded.

frobenius_norm_tf Not yet run
unverified (harvested, no recorded run)
PasqualeAuriemma/GCN-DIFFPOOL/models.py
code served (permissive licence) · get_code("0b25aff2096336db")
How it ran

no invocation recorded.

gen_ba Not yet run
unverified (harvested, no recorded run)
RexYing/diffpool/gen/data.py
code served (permissive licence) · get_code("854d177d079f1cd3")
How it ran

no invocation recorded.

gen_er Not yet run
unverified (harvested, no recorded run)
RexYing/diffpool/gen/data.py
code served (permissive licence) · get_code("9c9fec82cf325ebc")
How it ran

no invocation recorded.

kruskal Not yet run
unverified (harvested, no recorded run)
RexYing/diffpool/partition.py
code served (permissive licence) · get_code("eb9659a87e71ae20")
How it ran

no invocation recorded.

node_dict Not yet run
unverified (harvested, no recorded run)
RexYing/diffpool/util.py
code served (permissive licence) · get_code("d3550a1e301975aa")
How it ran

no invocation recorded.

node_iter Not yet run
unverified (harvested, no recorded run)
RexYing/diffpool/util.py
code served (permissive licence) · get_code("f55bf28721abc66d")
How it ran

no invocation recorded.

train Not yet run
unverified (harvested, no recorded run)
basiralab/reproduciblefedgnn/federated_reproducibility/main_gcn.py
code served (permissive licence) · get_code("640adefa68f0073d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 later) and 18 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.

FunctionAlso present inEarliest-datedOthers named
exp_moving_avg 2 other papers
2 later
1903.03894 (2019-03)
GNNExplainer: Generating Explanations for Graph Neural Networks
2407.01979 (2024-07)
Unveiling Global Interactive Patterns across Graphs: Towards Interpretable Graph Neural Networks
kruskal 1 other paper
1 later
2407.01979 (2024-07)
Unveiling Global Interactive Patterns across Graphs: Towards Interpretable Graph Neural Networks

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 2026-11-21.

Repositories linked to this paper

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

Abstract

Recently, graph neural networks (GNNs) have revolutionized the field of graph representation learning through effectively learned node embeddings, and achieved state-of-the-art results in tasks such as node classification and link prediction. However, current GNN methods are inherently flat and do not learn hierarchical representations of graphs---a limitation that is especially problematic for the task of graph classification, where the goal is to predict the label associated with an entire graph. Here we propose DiffPool, a differentiable graph pooling module that can generate hierarchical representations of graphs and can be combined with various graph neural network architectures in an end-to-end fashion. DiffPool learns a differentiable soft cluster assignment for nodes at each layer of a deep GNN, mapping nodes to a set of clusters, which then form the coarsened input for the next GNN layer. Our experimental results show that combining existing GNN methods with DiffPool yields an average improvement of 5-10% accuracy on graph classification benchmarks, compared to all existing pooling approaches, achieving a new state-of-the-art on four out of five benchmark data sets.

For agents

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

get_harvested_code_for_paper("1806.08804")
get_code_for_paper("1806.08804")
have("1806.08804")

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

curl https://syntology.ai/api/ran/1806.08804.json

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

[![Syntology run record](https://syntology.ai/api/ran/1806.08804.svg)](https://syntology.ai/paper/1806.08804)

Connect an agent — have() is free.