SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.16618 · ICLR · 2023

Confidence-Based Feature Imputation for Graphs with Partially Known Features

Jin Choi, Daeho Um, Jiwoong Park, Seulki Park

arXiv · PDF · 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; 0 honoured the contract we drafted, 1 violated it and 5 have no contract verdict; 4 failures of our instrument; 2 no recorded run), measured 2026-11-21; 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/2305.16618.json · badge: /api/ran/2305.16618.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
daehoum1/pcfi canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
FunctionStatusWhere it lives
get_conv Ran
ran; our contract draft was wrong, not the code
daehoum1/pcfi/models.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ssjcode/FSD-CAP, is pointer only (licence: NONE) · get_code("0cc5f5182936b878")
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 ssjcode/FSD-CAP, 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.

get_mask Ran
ran; our contract draft was wrong, not the code
daehoum1/pcfi/utils.py
code served (permissive licence) · get_code("12003dcf4549f5a3")
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 1, dtype family float, shape (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 torch

torch.manual_seed(0)

# The function takes:
# - idx: a tensor of indices
# - num_nodes: an integer representing the total number of nodes

# Create a small tensor of indices
idx = torch.tensor([0, 2, 4])

# Create a num_nodes value
num_nodes = 6

ARGS = {
    "idx": idx,
    "num_nodes": num_nodes,
}
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 1, dtype family float, shape (6,) — shape (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
get_missing_feature_mask Ran
ran, violated the contract we drafted
daehoum1/pcfi/utils.py
code served (permissive licence) · get_code("1c865fe69ab38ec4")
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 (4, 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 torch

torch.manual_seed(0)

ARGS = {
    'rate': 0.3,
    'n_nodes': 4,
    'n_features': 5,
    'seed': 42,
    'type': 'uniform'
}
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 (4, 5) — shape (4, 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
get_model Ran
ran on a synthesized input
daehoum1/pcfi/models.py
code served (permissive licence) · get_code("37e7b0e5f0549df5")
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 GNN. 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
from torch_geometric.nn import GCNConv

torch.manual_seed(0)

# Create a small stand-in class for args
class Args:
    def __init__(self):
        self.num_layers = 3
        self.hidden_dim = 8
        self.dropout = 0.1
        self.jk = False

# Create minimal tensors
num_features = 4
num_classes = 3
num_nodes = 6

x = torch.randn(num_nodes, num_features)
edge_index = torch.tensor([
    [0, 1, 2, 3, 4, 5],
    [1, 2, 3, 4, 5, 0]
], dtype=torch.long)

mask = torch.tensor([True, True, False, False, True, True])

ARGS = {
    'model_name': 'gcn',
    'num_features': num_features,
    'num_classes': num_classes,
    'edge_index': edge_index,
    'x': x,
    'args': Args(),
    'mask': mask
}
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 GNN — a GNN output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a GNN 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
inference_full_batch Ran
ran; our contract draft was wrong, not the code
daehoum1/pcfi/evaluation.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ssjcode/FSD-CAP, is pointer only (licence: NONE) · get_code("09a790aab330a10e")
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 ssjcode/FSD-CAP, 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.

inference_sampled Ran
ran on a synthesized input
daehoum1/pcfi/evaluation.py
code served (permissive licence) · get_code("17f844b866904bc9")
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, 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

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def inference(self, x, inference_loader, device):
        # Simple inference method that uses all three parameters
        output = self.linear(x)
        for batch in inference_loader:
            if isinstance(batch, (tuple, list)):
                batch_x, batch_y = batch
            else:
                batch_x = batch
            output = output + self.linear(batch_x)
        return output.to(device)

model = SimpleModel()
x = torch.randn(2, 4)
inference_loader = [
    (torch.randn(2, 4), torch.randn(2, 3)),
    (torch.randn(2, 4), torch.randn(2, 3))
]
device = torch.device('cpu')

ARGS = {
    'model': model,
    'x': x,
    'inference_loader': inference_loader,
    'device': device
}
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, 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: none by the harness; any seed is in the setup block
keep_only_largest_connected_component Ran
ran; our contract draft was wrong, not the code
daehoum1/pcfi/data_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ssjcode/FSD-CAP, is pointer only (licence: NONE) · get_code("9e01cbb3f563b6df")
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 ssjcode/FSD-CAP, 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.

test Ran
ran on a synthesized input
daehoum1/pcfi/evaluation.py
code served (permissive licence) · get_code("e9b81d3e51978cf1")
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. 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

torch.manual_seed(0)

# Build a minimal GNN model that takes x and edge_index
class SimpleGNN(nn.Module):
    def __init__(self, in_features, out_features):
        super().__init__()
        self.linear = nn.Linear(in_features, out_features)
    
    def forward(self, x, edge_index):
        # Just apply a linear layer, ignore edge_index for simplicity
        return self.linear(x)

# Create model
model = SimpleGNN(in_features=4, out_features=3)
model.eval()

# Create node features (2 nodes, 4 features each)
x = torch.randn(2, 4)

# Create edge_index (COO format)
edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long).t().contiguous()

# Create a minimal data object with required attributes
class Data:
    def __init__(self):
        self.edge_index = edge_index
        self.y = torch.tensor([0, 1], dtype=torch.long)
        self.train_mask = torch.tensor([True, False])
        self.val_mask = torch.tensor([False, True])
        self.test_mask = torch.tensor([False, True])
    
    def __call__(self, *mask_names):
        # Returns tuples of (name, mask) for each mask name
        masks = {
            'train_mask': self.train_mask,
            'val_mask': self.val_mask,
            'test_mask': self.test_mask,
        }
        return [(name, masks[name]) for name in mask_names]

data = Data()

# Create a minimal evaluator (optional, can be None)
class Evaluator:
    def eval(self, input_dict):
        return {'acc': 0.8}

evaluator = Evaluator()

ARGS = {
    'model': model,
    'x': x,
    'data': data,
    'logits': None,
    'evaluator': evaluator,
    'inference_loader': None,
    'device': 'cpu'
}
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
test Ran
ran on a synthesized input
daehoum1/pcfi/utils_link.py
code served (permissive licence) · get_code("e2d3e844742319c9")
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

torch.manual_seed(0)

class SimpleGNNModel(nn.Module):
    def __init__(self, input_dim, hidden_dim):
        super().__init__()
        self.linear = nn.Linear(input_dim, hidden_dim)
    
    def encode(self, x, edge_index):
        # Simple encoding: just pass through linear layer
        return self.linear(x)
    
    def test(self, z, pos_edge_index, neg_edge_index):
        # Simple test method: compute some metric
        # z shape: (num_nodes, hidden_dim)
        # pos_edge_index shape: (2, num_pos_edges)
        # neg_edge_index shape: (2, num_neg_edges)
        
        # Extract node embeddings for positive edges
        pos_src = z[pos_edge_index[0]]
        pos_dst = z[pos_edge_index[1]]
        pos_score = (pos_src * pos_dst).sum(dim=1)
        
        # Extract node embeddings for negative edges
        neg_src = z[neg_edge_index[0]]
        neg_dst = z[neg_edge_index[1]]
        neg_score = (neg_src * neg_dst).sum(dim=1)
        
        # Return some metric (e.g., mean scores)
        return pos_score.mean() - neg_score.mean()

# Create model
model = SimpleGNNModel(input_dim=4, hidden_dim=3)

# Create node features: 2 nodes, 4 features each
x = torch.randn(2, 4)

# Create training edge index: (2, num_edges)
train_pos_edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long)

# Create positive test edge index
pos_edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long)

# Create negative test edge index
neg_edge_index = torch.tensor([[0, 0], [1, 1]], dtype=torch.long)

ARGS = {
    "model": model,
    "x": x,
    "train_pos_edge_index": train_pos_edge_index,
    "pos_edge_index": pos_edge_index,
    "neg_edge_index": neg_edge_index,
}
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
train Ran
ran on a synthesized input
daehoum1/pcfi/utils_link.py
code served (permissive licence) · get_code("40895c9e127bcae7")
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

torch.manual_seed(0)

class SimpleVAE(nn.Module):
    def __init__(self, input_dim, latent_dim):
        super().__init__()
        self.input_dim = input_dim
        self.latent_dim = latent_dim
        self.encoder_linear = nn.Linear(input_dim, latent_dim)
        self.decoder_linear = nn.Linear(latent_dim, input_dim)
        
    def encode(self, x, edge_index):
        # Use x and edge_index to produce latent representation
        z = self.encoder_linear(x)
        return z
    
    def recon_loss(self, z, edge_index):
        # Reconstruct and compute loss
        recon = self.decoder_linear(z)
        loss = torch.nn.functional.mse_loss(recon, torch.randn_like(recon))
        return loss

# Create model
model = SimpleVAE(input_dim=4, latent_dim=3)

# Create small tensors
x = torch.randn(2, 4)  # 2 nodes, 4 features
train_pos_edge_index = torch.tensor([[0, 1], [1, 0]], dtype=torch.long).t().contiguous()  # 2x2 edge index

# Create optimizer
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)

ARGS = {
    "model": model,
    "x": x,
    "train_pos_edge_index": train_pos_edge_index,
    "optimizer": optimizer
}
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 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
get_component Not yet run
unverified (harvested, no recorded run)
daehoum1/pcfi/data_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ssjcode/FSD-CAP, is pointer only (licence: NONE) · get_code("c68140584cc0e991")
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 ssjcode/FSD-CAP, 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. No invocation recorded.

get_largest_connected_component Not yet run
unverified (harvested, no recorded run)
daehoum1/pcfi/data_utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ssjcode/FSD-CAP, is pointer only (licence: NONE) · get_code("b36878b25e3455a3")
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 ssjcode/FSD-CAP, 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. 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, 8 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 7 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.

FunctionAlso present inEarliest-datedOthers named
get_component 7 other papers
6 dated earlier, 1 later
1911.05485 (2019-11)
Diffusion Improves Graph Learning
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features
2206.10991 (2022-06)
Understanding convolution on graphs via energies
and 4 more not named here
get_largest_connected_component 6 other papers
5 dated earlier, 1 later
1911.05485 (2019-11)
Diffusion Improves Graph Learning
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features
2206.10991 (2022-06)
Understanding convolution on graphs via energies
and 3 more not named here
get_mask 2 other papers
1 dated earlier, 1 later
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features
2601.18938 (2026-01)
FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation
inference_full_batch 2 other papers
1 dated earlier, 1 later
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features
2601.18938 (2026-01)
FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation
keep_only_largest_connected_component 2 other papers
1 dated earlier, 1 later
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features
2601.18938 (2026-01)
FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation
get_conv 1 other paper
1 later
2601.18938 (2026-01)
FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation
get_missing_feature_mask 1 other paper
1 later
2601.18938 (2026-01)
FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation
inference_sampled 1 other paper
1 dated earlier
2111.12128 (2021-11)
On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features

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

This paper investigates a missing feature imputation problem for graph learning tasks. Several methods have previously addressed learning tasks on graphs with missing features. However, in cases of high rates of missing features, they were unable to avoid significant performance degradation. To overcome this limitation, we introduce a novel concept of channel-wise confidence in a node feature, which is assigned to each imputed channel feature of a node for reflecting certainty of the imputation. We then design pseudo-confidence using the channel-wise shortest path distance between a missing-feature node and its nearest known-feature node to replace unavailable true confidence in an actual learning process. Based on the pseudo-confidence, we propose a novel feature imputation scheme that performs channel-wise inter-node diffusion and node-wise inter-channel propagation. The scheme can endure even at an exceedingly high missing rate (e.g., 99.5%) and it achieves state-of-the-art accuracy for both semi-supervised node classification and link prediction on various datasets containing a high rate of missing features. Codes are available at https://github.com/daehoum1/pcfi.

For agents

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

get_harvested_code_for_paper("2305.16618")
get_code_for_paper("2305.16618")
have("2305.16618")

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

curl https://syntology.ai/api/ran/2305.16618.json

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

[![Syntology run record](https://syntology.ai/api/ran/2305.16618.svg)](https://syntology.ai/paper/2305.16618)

Connect an agent — have() is free.