SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.05718 · NeurIPS · 2020

Principal Neighbourhood Aggregation for Graph Nets

Pietro Liò, Gabriele Corso, Luca Cavalleri, Dominique Beaini, Petar Veličković

arXiv · PDF · Open in the Atlas

Code that ran

33 of 55 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 4 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 22 failures of our instrument; 22 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/2004.05718.json · badge: /api/ran/2004.05718.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
changminwu/expandergnn community
Papers with Code archive
18 ran · 3 no recorded run
Saro00/DGN community
Papers with Code archive
12 ran · 14 no recorded run
cvignac/SMP community
Papers with Code archive
3 ran · 4 no recorded run
alexOarga/haiku-geometric community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
ExpanderLinear Ran
ran on a synthesized input · constructed an object; no result was computed
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("bbcfddab0672587b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

ExpanderLinearFunction Ran
ran on a synthesized input · constructed an object; no result was computed
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("4aecc0c5ee5d5919")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

ExpanderScatterLinear Ran
ran on a synthesized input · constructed an object; no result was computed
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("528a64a346d98ccd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

Linear Ran
ran on a synthesized input
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("e511d58ec99ede06")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import torch.nn as nn
from torch import Tensor
from torch.nn.init import _calculate_correct_fan, calculate_gain
import math

def kaiming_init_with_gain(x: Tensor, gain: float, a=0, mode='fan_in', nonlinearity='relu'):
    fan = _calculate_correct_fan(x, mode)
    non_linearity_gain = calculate_gain(nonlinearity, a)
    std = non_linearity_gain / math.sqrt(fan)
    bound = math.sqrt(3.0) * std * gain
    with torch.no_grad():
        return x.uniform_(-bound, bound)

class Linear(nn.Module):
    """ Linear layer with potentially smaller parameters at initialization. """
    __constants__ = ['bias', 'in_features', 'out_features']

    def __init__(self, in_features, out_features, bias=True, gain: float=1.0):
        super().__init__()
        self.gain = gain
        self.lin = nn.Linear(in_features, out_features, bias)

    def reset_parameters(self):
        kaiming_init_with_gain(self.lin.weight, self.gain)
        if self.lin.bias is not None:
            nn.init.normal_(self.lin.bias, 0, self.gain / math.sqrt(self.lin.out_features))

    def forward(self, x):
        return self.lin.forward(x)


INIT_ARGS = {
    'in_features': 4,
    'out_features': 3,
    'bias': True,
    'gain': 0.5
}

ARGS = {
    'x': torch.randn(2, 4)
}

if __name__ == '__main__':
    layer = Linear(**INIT_ARGS)
    layer.reset_parameters()
    output = layer(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 (2, 3) — shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
LinearLayer Ran
ran on a synthesized input · constructed an object; no result was computed
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("6f4fac2426ac1ca3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

MLP Ran
ran on a synthesized input
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("5fc1afe08991ab11")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

XtoX Ran
ran on a synthesized input
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("ef6906c0d3364ced")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import torch.nn as nn
from torch import Tensor
from torch.nn.init import _calculate_correct_fan, calculate_gain
import math

class Linear(nn.Module):
    """ Linear layer with potentially smaller parameters at initialization. """
    __constants__ = ['bias', 'in_features', 'out_features']

    def __init__(self, in_features, out_features, bias=True, gain: float=1.0):
        super().__init__()
        self.gain = gain
        self.lin = nn.Linear(in_features, out_features, bias)

    def reset_parameters(self):
        kaiming_init_with_gain(self.lin.weight, self.gain)
        if self.lin.bias is not None:
            nn.init.normal_(self.lin.bias, 0, self.gain / math.sqrt(self.lin.out_features))

    def forward(self, x):
        return self.lin.forward(x)

def kaiming_init_with_gain(x: Tensor, gain: float, a=0, mode='fan_in', nonlinearity='relu'):
    fan = _calculate_correct_fan(x, mode)
    non_linearity_gain = calculate_gain(nonlinearity, a)
    std = non_linearity_gain / math.sqrt(fan)
    bound = math.sqrt(3.0) * std * gain
    with torch.no_grad():
        return x.uniform_(-bound, bound)

class XtoX(Linear):

    def forward(self, x, batch_info: dict=None):
        return self.lin.forward(x)


INIT_ARGS = {
    'in_features': 4,
    'out_features': 3,
    'bias': True,
    'gain': 1.0,
}

ARGS = {
    'x': torch.randn(2, 4),
    'batch_info': None,
}

if __name__ == '__main__':
    model = XtoX(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 (2, 3) — shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
aggregate_identity Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("a03e9cc59c2671f7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

aggregate_max Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("6a8ff41179e540bd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

aggregate_max Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("0cb93154c335108f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_mean Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("b19d1fdc25bad3bd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_min Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("3b1fc26527b0808f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

aggregate_min Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("ff879af48e34176e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_moment Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("481088b2fcfe7538")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_moment_3 Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("12e2cedd3aa5d6ee")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_moment_4 Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("1459f6559164001b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_moment_5 Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("4a7092a5a3c9f554")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_normalised_mean Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("7e7c9a7e4e3f56ed")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

aggregate_std Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("6eb59d131c4ff569")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_sum Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("5aeed6a78e303573")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

aggregate_var Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("92fa1e926f5bd985")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

get_activation Ran
ran; our contract draft was wrong, not the code
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("fe76e55f0b804b7b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

get_adjacency_from_gradient_of_features Ran
ran; our contract draft was wrong, not the code
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("00473157272f2713")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

get_eig_adjacency Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("b4f4cbb16613e6d5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

kaiming_init_with_gain Ran
ran; our contract draft was wrong, not the code
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("244b5dd39c13f84a")
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 Tensor, 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 torch
from torch import Tensor
import math

torch.manual_seed(0)

# Create a small tensor with shape (2, 4) for testing
# The function uses this tensor with _calculate_correct_fan and uniform_
x = torch.randn(2, 4)

# gain is a float multiplier
gain = 1.5

# a is the negative slope for leaky_relu (default 0 for relu)
a = 0.01

# mode determines how fan is calculated ('fan_in' or 'fan_out')
mode = 'fan_in'

# nonlinearity type for calculate_gain
nonlinearity = 'relu'

ARGS = {
    'x': x,
    'gain': gain,
    'a': a,
    'mode': mode,
    'nonlinearity': nonlinearity
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 (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
sampler Ran
ran; our contract draft was wrong, not the code
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("ea989e2ffeb38bb6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

scale_amplification Ran
ran; our contract draft was wrong, not the code
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("1d260f12223b0d23")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

scale_amplification Ran
ran, honoured the contract we drafted
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("5d07dbe669b9f73a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

scale_attenuation Ran
ran; our contract draft was wrong, not the code
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("e1422930e833ac1b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

scale_attenuation Ran
ran, honoured the contract we drafted
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("d0d453f04b18e8f1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

scale_identity Ran
ran, honoured the contract we drafted
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("3d28677798f20e1f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

scale_identity Ran
ran, honoured the contract we drafted
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("2d32c33e16bad56e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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.

scale_linear Ran
ran; our contract draft was wrong, not the code
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("b673916e3a39dd5d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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.

DGNLayer Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("e4b24e60ca0d9599")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

DGNTower Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("ce5e689ba09ce45d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

EntryWiseX Not yet run
unverified (harvested, no recorded run)
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("8e16971a959b54a3")
How it ran

no invocation recorded.

EntrywiseU Not yet run
unverified (harvested, no recorded run)
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("9df9644a36246a29")
How it ran

no invocation recorded.

FCLayer Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("3aa8e705794c9d66")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

MultiLinearLayer Not yet run
unverified (harvested, no recorded run)
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("71500fc15d19f2bc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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. No invocation recorded.

PNAConv Not yet run
unverified (harvested, no recorded run)
alexOarga/haiku-geometric/haiku_geometric/nn/conv/pna_conv.py
pointer only (licence: NONE) · get_code("f610d8297c271f3c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexOarga/haiku-geometric 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. No invocation recorded.

PNALayer Not yet run
unverified (harvested, no recorded run)
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("e8010b28fdc39a80")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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. No invocation recorded.

PNATower Not yet run
unverified (harvested, no recorded run)
changminwu/expandergnn/layers/pna_layer.py
pointer only (licence: NONE) · get_code("2c7376a202ec32d2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, changminwu/expandergnn 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. No invocation recorded.

SMPLayer Not yet run
unverified (harvested, no recorded run)
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("789af8973c8347ee")
How it ran

no invocation recorded.

UtoU Not yet run
unverified (harvested, no recorded run)
cvignac/SMP/models/smp_layers.py
code served (permissive licence) · get_code("16f7d89050a5bb58")
How it ran

no invocation recorded.

aggregate_dirs Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("aeafec5c97bd497f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_mean Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("d7ead85a1d37c169")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_mean_amplified Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("be921ec112433f72")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_mean_attenuated Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("89ec98a7ccf842eb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_moment_rooted Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("cfb905cff1a241fd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_softmax Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("94492fb6844ac8f1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_softmin Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("3d7c8bb83a9f2a81")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_std Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("d38bd6844f4c5c7e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_sum Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("c6d0524ad9a473ce")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

aggregate_var Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("1ad750797e80563a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

scale_inverse_linear Not yet run
unverified (harvested, no recorded run)
Saro00/DGN/models/pytorch/dgn_layer.py
pointer only (licence: NONE) · get_code("c53e8d8bb36479c1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Saro00/DGN 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. No invocation recorded.

Also present in other papers' code

Of 55 samples lifted from this paper's linked repositories, 55 carry a function identity and 0 do not. Of the 55 distinct functions with an identity, 8 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 8 later) and 47 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
aggregate_max
brick of this paper's assembly: helper
2 other papers
2 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2310.18716 (2023-10)
Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding
aggregate_mean
brick of this paper's assembly: helper
2 other papers
2 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2310.18716 (2023-10)
Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding
aggregate_min
brick of this paper's assembly: helper
2 other papers
2 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2310.18716 (2023-10)
Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding
get_activation
brick of this paper's assembly: helper
2 other papers
2 later
2010.02863 (2020-10)
Directional Graph Networks
2409.19414 (2024-09)
Sequential Signal Mixing Aggregation for Message Passing Graph Neural Networks
kaiming_init_with_gain
brick of this paper's assembly: helper
1 other paper
1 later
2206.11010 (2022-06)
Agent-based Graph Neural Networks
scale_amplification
brick of this paper's assembly: helper
1 other paper
1 later
2010.02863 (2020-10)
Directional Graph Networks
scale_attenuation
brick of this paper's assembly: helper
1 other paper
1 later
2010.02863 (2020-10)
Directional Graph Networks
scale_identity
brick of this paper's assembly: helper
1 other paper
1 later
2010.02863 (2020-10)
Directional Graph 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

Graph Neural Networks (GNNs) have been shown to be effective models for different predictive tasks on graph-structured data. Recent work on their expressive power has focused on isomorphism tasks and countable feature spaces. We extend this theoretical framework to include continuous features-which occur regularly in real-world input domains and within the hidden layers of GNNs-and we demonstrate the requirement for multiple aggregation functions in this context. Accordingly, we propose Principal Neighbourhood Aggregation (PNA), a novel architecture combining multiple aggregators with degree-scalers (which generalize the sum aggregator). Finally, we compare the capacity of different models to capture and exploit the graph structure via a novel benchmark containing multiple tasks taken from classical graph theory, alongside existing benchmarks from realworld domains, all of which demonstrate the strength of our model. With this work, we hope to steer some of the GNN research towards new aggregation methods which we believe are essential in the search for powerful and robust models.

For agents

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

get_harvested_code_for_paper("2004.05718")
get_code_for_paper("2004.05718")
have("2004.05718")

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

curl https://syntology.ai/api/ran/2004.05718.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.05718.svg)](https://syntology.ai/paper/2004.05718)

Connect an agent — have() is free.