SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1905.09550 · 2019

Revisiting Graph Neural Networks: All We Have is Low-Pass Filters

arXiv · PDF · Open in the Atlas

Code that ran

17 of 18 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, 1 violated it and 11 have no contract verdict; 4 failures of our instrument; 1 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/1905.09550.json · badge: /api/ran/1905.09550.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
gear/gfnn canonical (classifier label)
classifier basis: census
12 ran · 0 no recorded run
hazdzz/gfNN reimplementation (classifier label)
classifier basis: classifier_2025
5 ran · 1 no recorded run
FunctionStatusWhere it lives
abs_rw_normalized_adjacency Ran
ran on a synthesized input
gear/gfnn/normalization.py
code served (permissive licence) · get_code("6f5f73e918b46e6f")
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 coo_matrix. 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.sparse as sp

# Create a small sparse adjacency matrix
# Using a 4x4 matrix for small size
rows = np.array([0, 1, 2, 3, 0, 2])
cols = np.array([1, 2, 3, 0, 3, 1])
data = np.array([1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
adj = sp.csr_matrix((data, (rows, cols)), shape=(4, 4))

ARGS = {
    "adj": adj
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type coo_matrix — a coo_matrix output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a coo_matrix 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
accuracy Ran
ran, honoured the contract we drafted
gear/gfnn/metrics.py
pointer only (licence: MIT); the stored copy get_code serves, from TianxiangZhao/GraphSmote, is pointer only (licence: NONE) · get_code("18557871f0c93b3f")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from TianxiangZhao/GraphSmote, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

aug_normalized_adjacency Ran
ran on a synthesized input
gear/gfnn/normalization.py
code served (permissive licence) · get_code("a4a8c0e8d8c8a86e")
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 coo_matrix. 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.sparse as sp

# Create a small sparse adjacency matrix
# Using a 4x4 matrix for small size
n_nodes = 4
# Create a simple sparse matrix (not too dense)
row = np.array([0, 1, 2, 3, 0, 2])
col = np.array([1, 2, 3, 0, 3, 1])
data = np.array([1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
adj = sp.coo_matrix((data, (row, col)), shape=(n_nodes, n_nodes))

ARGS = {
    "adj": adj
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type coo_matrix — a coo_matrix output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a coo_matrix 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
calc_gso Ran
ran; our contract draft was wrong, not the code
hazdzz/gfNN/script/utility.py
pointer only (licence: LGPL-2.1) · get_code("844b86547e59a9cb")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

calc_sgc_feature Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
hazdzz/gfNN/script/utility.py
pointer only (licence: LGPL-2.1) · get_code("4c178650ff16f272")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

cnv_sparse_mat_to_coo_tensor Ran
ran; our contract draft was wrong, not the code
hazdzz/gfNN/script/utility.py
pointer only (licence: LGPL-2.1) · get_code("03ccfc745e8f1df3")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

f1 Ran
ran on a synthesized input
gear/gfnn/metrics.py
code served (permissive licence) · get_code("cc1fd8ffd02dab36")
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. 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
from sklearn.metrics import f1_score

torch.manual_seed(0)

# Create a small batch of outputs and labels
batch_size = 2
num_classes = 4

# output: tensor that supports .max(1)[1] to get predictions
# Shape: (batch_size, num_classes)
output = torch.randn(batch_size, num_classes)

# labels: tensor that can be converted to numpy
# Shape: (batch_size,) with class indices from 0 to num_classes-1
labels = torch.tensor([0, 2], dtype=torch.long)

ARGS = {
    "output": output,
    "labels": labels,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, 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
fetch_normalization Ran
ran on a synthesized input
gear/gfnn/normalization.py
code served (permissive licence) · get_code("1a5b4299f5be0a79")
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 function. 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.sparse as sp

# Create a small sparse adjacency matrix
# 4x4 adjacency matrix with some edges
data = np.array([1.0, 1.0, 1.0, 1.0, 1.0, 1.0])
row = np.array([0, 0, 1, 1, 2, 3])
col = np.array([1, 2, 0, 3, 1, 2])
adj = sp.csr_matrix((data, (row, col)), shape=(4, 4))

ARGS = {
    'type': 'AugNormAdj',
    'alpha': 0.8,
    'scale': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
gaussian Ran
ran on a synthesized input
gear/gfnn/noise.py
code served (permissive licence) · get_code("5f4af3371e9442b6")
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

# Create a small numpy array to pass as 'f'
f = np.random.randn(2, 4)

# Define all parameters for the gaussian function
ARGS = {
    "f": f,
    "mean": 0.0,
    "std": 1.0,
    "add": False
}
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 (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
gaussian_mimic Ran
ran on a synthesized input
gear/gfnn/noise.py
code served (permissive licence) · get_code("62a1471638f46485")
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

# Create a small feature array with shape (2, 4)
# This will be used as input 'f' to the gaussian_mimic function
f = np.random.randn(2, 4).astype(np.float32)

# Create the ARGS dictionary with all parameters
ARGS = {
    'f': f,
    'add': False
}
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 (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
get_model Ran
ran on a synthesized input
gear/gfnn/models.py
code served (permissive licence) · get_code("d100375cd9ce3491")
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 GCN. 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 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 math

torch.manual_seed(0)

ARGS = {
    'model_opt': 'GCN',
    'nfeat': 4,
    'nclass': 3,
    'nhid': 5,
    'dropout': 0.2,
    'cuda': False,
    'degree': 2,
}
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 GCN — a GCN output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a GCN 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
load_webkb_data Ran
ran; our contract draft was wrong, not the code
hazdzz/gfNN/script/dataloader.py
pointer only (licence: LGPL-2.1) · get_code("9363b574f33a10a1")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

make_donuts Ran
ran on a synthesized input
gear/gfnn/synthetic_data.py
code served (permissive licence) · get_code("46ea2adafba4925f")
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 numpy as np
from sklearn import datasets
from sklearn.model_selection import ShuffleSplit
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import NearestNeighbors

# Set random seed for reproducibility
np.random.seed(0)

# Test the make_donuts function with minimal arguments
ARGS = {
    'n': 100,
    'noise': 0.2,
    'factor': 0.5,
    'test_size': 0.92,
    'nneigh': 5,
    'mesh': False,
    'mesh_step': 0.02,
}
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 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
make_graph Ran
ran on a synthesized input
gear/gfnn/synthetic_data.py
code served (permissive licence) · get_code("73b528f4485a27d4")
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 csr_matrix. 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
from sklearn.neighbors import NearestNeighbors

torch.manual_seed(0)

# Create a small sample dataset
# X should be a 2D array-like of shape (n_samples, n_features)
X = np.array([
    [0.0, 1.0, 2.0, 3.0],
    [1.0, 2.0, 3.0, 4.0],
    [2.0, 3.0, 4.0, 5.0],
    [3.0, 4.0, 5.0, 6.0],
    [4.0, 5.0, 6.0, 7.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float32)

ARGS = {
    'X': X,
    'n_neighbors': 4,
    'algo': 'ball_tree'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type csr_matrix — a csr_matrix output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a csr_matrix 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
norm_feat Ran
ran, violated the contract we drafted
hazdzz/gfNN/script/dataloader.py
pointer only (licence: LGPL-2.1) · get_code("54c8404c5585302e")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

superimpose_gaussian Ran
ran on a synthesized input
gear/gfnn/noise.py
code served (permissive licence) · get_code("c234ab2b90d78d36")
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, 6) 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

# Set seed for reproducibility
np.random.seed(0)

# Create a small feature matrix: 2 samples, 4 features
f = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float32)

ARGS = {
    "f": f,
    "k": 1.5
}
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 (2, 6) — shape (2, 6) 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
train_regression Ran
ran on a synthesized input
gear/gfnn/reddit.py
code served (permissive licence) · get_code("3d5322498444a628")
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. 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 torch.nn as nn
import torch.optim as optim

torch.manual_seed(0)

# Build a small model
class SimpleRegressor(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

model = SimpleRegressor()

# Create tiny train features and labels
train_features = torch.randn(2, 4)  # batch size 2, feature dim 4
train_labels = torch.randint(0, 3, (2,))  # batch size 2, 3 classes for cross_entropy

epochs = 2

ARGS = {
    "model": model,
    "train_features": train_features,
    "train_labels": train_labels,
    "epochs": epochs
}
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 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
load_citation_data Not yet run
unverified (harvested, no recorded run)
hazdzz/gfNN/script/dataloader.py
pointer only (licence: LGPL-2.1) · get_code("f5e957aeda88abd5")
How it ran

The body of this sample is withheld: LGPL-2.1, as GitHub detects it on this paper's repository, hazdzz/gfNN, is not in the permissive set Syntology serves inline. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 1 later) and 15 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
accuracy 14 other papers
1 dated earlier, 13 later
1902.07153 (2019-02)
Simplifying Graph Convolutional Networks
2002.07128 (2020-02)
Disease State Prediction From Single-Cell Data Using Graph Attention Networks
2103.08826 (2021-03)
GraphSMOTE: Imbalanced Node Classification on Graphs with Graph Neural Networks
and 11 more not named here
aug_normalized_adjacency 1 other paper
1 dated earlier
1902.07153 (2019-02)
Simplifying Graph Convolutional Networks
f1 1 other paper
1 dated earlier
1902.07153 (2019-02)
Simplifying Graph Convolutional 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 2026-11-22.

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 neural networks have become one of the most important techniques to solve machine learning problems on graph-structured data. Recent work on vertex classification proposed deep and distributed learning models to achieve high performance and scalability. However, we find that the feature vectors of benchmark datasets are already quite informative for the classification task, and the graph structure only provides a means to denoise the data. In this paper, we develop a theoretical framework based on graph signal processing for analyzing graph neural networks. Our results indicate that graph neural networks only perform low-pass filtering on feature vectors and do not have the non-linear manifold learning property. We further investigate their resilience to feature noise and propose some insights on GCN-based graph neural network design.

For agents

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

get_harvested_code_for_paper("1905.09550")
get_code_for_paper("1905.09550")
have("1905.09550")

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

curl https://syntology.ai/api/ran/1905.09550.json

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

[![Syntology run record](https://syntology.ai/api/ran/1905.09550.svg)](https://syntology.ai/paper/1905.09550)

Connect an agent — have() is free.