SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_29668 · AAAI · 2024

Robust Node Classification on Graph Data with Graph and Label Noise

Lei Feng, Michael Witbrock, Yang Chen, Zhenyun Deng, Yonghua Zhu, Robert Amor

Open in the Atlas

Code that ran

10 of 12 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, 0 violated it and 6 have no contract verdict; 3 failures of our instrument; 2 no recorded run), measured 2026-11-22; 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/aaai_29668.json · badge: /api/ran/aaai_29668.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
yhzhu66/RNCGLN canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
FunctionStatusWhere it lives
Ncontrast Ran
ran on a synthesized input
yhzhu66/RNCGLN/models/Semi_RNCGLN.py
code served (permissive licence) · get_code("2186d1f50390578d")
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

torch.manual_seed(0)

# Create small tensors for the function
batch_size = 2
num_features = 4

# x_dis: similarity/distance scores, shape (batch_size, batch_size)
x_dis = torch.randn(batch_size, batch_size)

# adj_label: adjacency label matrix, shape (batch_size, batch_size)
# Binary matrix indicating positive pairs
adj_label = torch.randint(0, 2, (batch_size, batch_size)).float()

# tau: temperature parameter (scalar)
tau = 1.0

# train_index_sort: indices for selecting which samples to use
# Shape should be compatible with indexing x_dis_sum_mid (which has batch_size elements)
train_index_sort = torch.tensor([0, 1])

ARGS = {
    "x_dis": x_dis,
    "adj_label": adj_label,
    "tau": tau,
    "train_index_sort": train_index_sort,
}
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 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
build_pair_p Ran
ran on a synthesized input
yhzhu66/RNCGLN/utils/Noise_about.py
code served (permissive licence) · get_code("7431b728d8362271")
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 (5, 5) 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

ARGS = {
    "size": 5,
    "noise": 0.3
}
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 (5, 5) — shape (5, 5) 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
build_uniform_P Ran
ran on a synthesized input
yhzhu66/RNCGLN/utils/Noise_about.py
code served (permissive licence) · get_code("99105e2dce5e045a")
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 (5, 5) 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

# This function takes two parameters: size (int) and noise (float)
# size: the number of classes (determines the shape of the noise matrix)
# noise: probability parameter between 0 and 1

ARGS = {
    "size": 5,
    "noise": 0.3
}
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 (5, 5) — shape (5, 5) 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
encode_onehot Ran
ran, honoured the contract we drafted
yhzhu66/RNCGLN/utils/process.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from zxlearningdeep/InfoMGF, is pointer only (licence: NONE) · get_code("6c897974e5ceb4aa")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from zxlearningdeep/InfoMGF, 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.

encode_onehot Ran
ran; our contract draft was wrong, not the code
yhzhu66/RNCGLN/utils/utils.py
code served (permissive licence) · get_code("996030d9ac2a825d")
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 (5, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# The function encode_onehot takes labels as input
# It calls labels.max() and uses labels as indices into an eye matrix
# So labels should be a numpy array of non-negative integers

labels = np.array([0, 2, 1, 0, 2], dtype=np.int64)

ARGS = {
    "labels": labels
}
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 (5, 3) — shape (5, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_A_r Ran
ran; our contract draft was wrong, not the code
yhzhu66/RNCGLN/models/Semi_RNCGLN.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from larryuestc/cocomg, is pointer only (licence: NONE) · get_code("3c418354518f2bfc")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from larryuestc/cocomg, 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.

preprocess Ran
ran on a synthesized input
yhzhu66/RNCGLN/utils/utils.py
code served (permissive licence) · get_code("408d0fc7e5274a43")
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
import scipy.sparse as sp
import torch

torch.manual_seed(0)

# Create a small sparse adjacency matrix (4x4)
n_nodes = 4
adj = sp.lil_matrix((n_nodes, n_nodes))
adj[0, 1] = 1
adj[1, 0] = 1
adj[1, 2] = 1
adj[2, 1] = 1
adj[2, 3] = 1
adj[3, 2] = 1
adj = adj.tocsr()

# Create a small sparse feature matrix (4x3)
n_features = 3
features = sp.lil_matrix((n_nodes, n_features))
features[0, 0] = 1.0
features[1, 1] = 1.0
features[2, 2] = 1.0
features[3, 0] = 0.5
features = features.tocsr()

# Create labels (one per node)
labels = np.array([0, 1, 0, 1])

ARGS = {
    'adj': adj,
    'features': features,
    'labels': labels,
    'preprocess_adj': False,
    'preprocess_feature': False,
    'sparse': False,
    'device': 'cpu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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
preprocess_features Ran
ran on a synthesized input
yhzhu66/RNCGLN/utils/process.py
code served (permissive licence) · get_code("0197b5d5b7be9292")
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 matrix, ndim 2, dtype family float, shape (4, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import scipy.sparse as sp

# Create a small sparse feature matrix
# The function expects a matrix that can be summed along axis 1
# and supports .dot() operations (scipy sparse matrix)
np.random.seed(0)

# Create a small sparse matrix: 4 samples, 3 features
# Using lil_matrix for efficient construction
features_dense = np.random.rand(4, 3).astype(np.float32)
# Convert to sparse format
features = sp.csr_matrix(features_dense)

ARGS = {
    "features": features,
    "eps": 1e-06
}
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 matrix, ndim 2, dtype family float, shape (4, 3) — shape (4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
row_normalize Ran
ran on a synthesized input
yhzhu66/RNCGLN/models/Semi_RNCGLN.py
code served (permissive licence) · get_code("8dd385911d9fd060")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 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

torch.manual_seed(0)

# The function expects A to be a tensor with:
# - A.sum(dim=-1) method (works on any tensor)
# - .clamp(min=0.0) method (tensor method)
# - .pow(-1) method (tensor method)
# - .unsqueeze(-1) method (tensor method)
# - element-wise multiplication with * operator

# Create a small dense matrix (batch_size=2, features=4)
A = torch.randn(2, 4)

ARGS = {
    "A": A
}
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 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
tensor2onehot Ran
ran; our contract draft was wrong, not the code
yhzhu66/RNCGLN/utils/utils.py
code served (permissive licence) · get_code("e8bec6780bc6ac76")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (6, 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

torch.manual_seed(0)

# Create a small tensor of class labels
# labels should be a 1D tensor with integer values from 0 to max_label
labels = torch.tensor([0, 1, 2, 1, 0, 3])

ARGS = {
    "labels": labels
}
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 Tensor, ndim 2, dtype family float, shape (6, 4) — shape (6, 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
load_single_graph Not yet run
unverified (harvested, no recorded run)
yhzhu66/RNCGLN/utils/process.py
code served (permissive licence) · get_code("f3bc0a2c3b86c529")
How it ran

no invocation recorded.

multiclass_noisify Not yet run
unverified (harvested, no recorded run)
yhzhu66/RNCGLN/utils/Noise_about.py
code served (permissive licence) · get_code("daf2a1f7a7323a3a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 5 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 0 later) and 7 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
encode_onehot 3 other papers
3 same year, order not determinable
2403.15520 (2024-03)
GTC: GNN-Transformer Co-contrastive Learning for Self-supervised Heterogeneous Graph Representation
2407.16863 (2024-07)
Balanced Multi-Relational Graph Clustering
2409.17386 (2024-09)
Beyond Redundancy: Information-aware Unsupervised Multiplex Graph Structure Learning
tensor2onehot 2 other papers
2 dated earlier
2005.06149 (2020-05)
DeepRobust: A PyTorch Library for Adversarial Attacks and Defenses
2306.09104 (2023-06)
On Strengthening and Defending Graph Reconstruction Attack with Markov Chain Approximation
encode_onehot 1 other paper
1 dated earlier
2005.06149 (2020-05)
DeepRobust: A PyTorch Library for Adversarial Attacks and Defenses
get_A_r 1 other paper
1 dated earlier
2308.01606 (2023-08)
Unsupervised Multiplex Graph Learning with Complementary and Consistent Information
multiclass_noisify 1 other paper
1 dated earlier
2006.13554 (2020-06)
Normalized Loss Functions for Deep Learning with Noisy Labels

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

Current research for node classification focuses on dealing with either graph noise or label noise, but few studies consider both of them. In this paper, we propose a new robust node classification method to simultaneously deal with graph noise and label noise. To do this, we design a graph contrastive loss to conduct local graph learning and employ selfattention to conduct global graph learning. They enable us to improve the expressiveness of node representation by using comprehensive information among nodes. We also utilize pseudo graphs and pseudo labels to deal with graph noise and label noise, respectively. Furthermore, We numerically validate the superiority of our method in terms of robust node classification compared with all comparison methods.

For agents

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

get_harvested_code_for_paper("aaai_29668")
get_code_for_paper("aaai_29668")
have("aaai_29668")

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

curl https://syntology.ai/api/ran/aaai_29668.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_29668.svg)](https://syntology.ai/paper/aaai_29668)

Connect an agent — have() is free.