SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.03473 · 2022

Higher-order Clustering and Pooling for Graph Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 1 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/2209.03473.json · badge: /api/ran/2209.03473.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
alexduvalinho/hoscpool canonical (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
FunctionStatusWhere it lives
dense_hoscpool Ran
ran on a synthesized input
alexduvalinho/hoscpool/hoscpool.py
code served (permissive licence) · get_code("45d6fd62d92c0c55")
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)

# Tiny batch setup: batch_size=2, num_nodes=4, feature_dim=3, num_clusters=2
batch_size = 2
num_nodes = 4
feature_dim = 3
num_clusters = 2

# Node feature tensor: (batch_size, num_nodes, feature_dim)
x = torch.randn(batch_size, num_nodes, feature_dim)

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

# Cluster assignment matrix: (batch_size, num_nodes, num_clusters)
s = torch.randn(batch_size, num_nodes, num_clusters)

# Optional parameters
mu = 0.1
alpha = 0.5
new_ortho = False
mask = None

ARGS = {
    "x": x,
    "adj": adj,
    "s": s,
    "mu": mu,
    "alpha": alpha,
    "new_ortho": new_ortho,
    "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
evaluate Ran
ran on a synthesized input
alexduvalinho/hoscpool/example.py
code served (permissive licence) · get_code("c08e1f7696006e20")
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 float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from torch_geometric.data import Data, DataLoader

torch.manual_seed(0)

class SimpleGNNModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.lin = nn.Linear(4, 3)
    
    def forward(self, x, edge_index, batch):
        # Model must return 4 values: out, mc, o, s
        # out should have shape (num_nodes, num_classes) for argmax and eq operations
        out = self.lin(x)
        mc = torch.tensor(0.0)
        o = torch.tensor(0.0)
        s = torch.tensor(0.0)
        return out, mc, o, s

# Create tiny synthetic graph data
# Create 2 graph samples with 3 nodes each
x1 = torch.randn(3, 4)  # 3 nodes, 4 features
edge_index1 = torch.tensor([[0, 1, 1, 2], [1, 0, 2, 1]], dtype=torch.long)
y1 = torch.tensor([0, 1, 0], dtype=torch.long)
data1 = Data(x=x1, edge_index=edge_index1, y=y1, batch=torch.tensor([0, 0, 0]))

x2 = torch.randn(3, 4)  # 3 nodes, 4 features
edge_index2 = torch.tensor([[0, 1, 2], [1, 2, 0]], dtype=torch.long)
y2 = torch.tensor([1, 0, 1], dtype=torch.long)
data2 = Data(x=x2, edge_index=edge_index2, y=y2, batch=torch.tensor([1, 1, 1]))

# Create a simple loader (list of Data objects works as iterable)
loader = [data1, data2]
# Add len and dataset attributes for compatibility
loader_obj = type('SimpleLoader', (), {
    '__iter__': lambda self: iter(loader),
    '__len__': lambda self: len(loader),
    'dataset': type('Dataset', (), {'__len__': lambda s: sum(len(d.y) for d in loader)})()
})()

model = SimpleGNNModel()

ARGS = {
    'model': model,
    'loader': loader_obj,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, 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
train Ran
ran on a synthesized input
alexduvalinho/hoscpool/example.py
code served (permissive licence) · get_code("526947f58a277cda")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from torch_geometric.data import Batch, Data

torch.manual_seed(0)

# Define a minimal model that matches the function's usage
class MinimalGNNModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 2)
    
    def forward(self, x, edge_index, batch):
        # Returns out, mc, o, s as expected by the train function
        out = self.linear(x)
        mc = torch.tensor(0.1)  # regularization term
        o = torch.tensor(0.05)  # another regularization term
        s = torch.tensor(0.02)  # unused in train but returned
        return out, mc, o, s

# Define a minimal criterion (loss function)
class MinimalCriterion(nn.Module):
    def forward(self, out, y):
        return torch.nn.functional.mse_loss(out, y)

# Create minimal batch data
# x: node features (2 nodes, 4 features each)
x = torch.randn(2, 4)
# edge_index: edge connectivity (2 nodes connected)
edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long).t().contiguous()
# y: target labels (2 nodes, 2 classes)
y = torch.randn(2, 2)
# batch: batch assignment (both nodes in batch 0)
batch_assignment = torch.tensor([0, 0], dtype=torch.long)

# Create Data objects and a loader
data = Data(x=x, edge_index=edge_index, y=y, batch=batch_assignment)
loader = [data]  # Simple list loader with one batch

# Create model, optimizer, and criterion
model = MinimalGNNModel()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)
criterion = MinimalCriterion()

ARGS = {
    "loader": loader,
    "optimizer": optimizer,
    "criterion": criterion,
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torch_geometric — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, 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
get_data Not yet run
unverified (harvested, no recorded run)
alexduvalinho/hoscpool/example.py
code served (permissive licence) · get_code("8c06449ddef1778b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 4 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

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 achieve state-of-the-art performance on a plethora of graph classification tasks, especially due to pooling operators, which aggregate learned node embeddings hierarchically into a final graph representation. However, they are not only questioned by recent work showing on par performance with random pooling, but also ignore completely higher-order connectivity patterns. To tackle this issue, we propose HoscPool, a clustering-based graph pooling operator that captures higher-order information hierarchically, leading to richer graph representations. In fact, we learn a probabilistic cluster assignment matrix end-to-end by minimising relaxed formulations of motif spectral clustering in our objective function, and we then extend it to a pooling operator. We evaluate HoscPool on graph classification tasks and its clustering component on graphs with ground-truth community structure, achieving best performance. Lastly, we provide a deep empirical analysis of pooling operators' inner functioning.

For agents

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

get_harvested_code_for_paper("2209.03473")
get_code_for_paper("2209.03473")
have("2209.03473")

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

curl https://syntology.ai/api/ran/2209.03473.json

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

[![Syntology run record](https://syntology.ai/api/ran/2209.03473.svg)](https://syntology.ai/paper/2209.03473)

Connect an agent — have() is free.