SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.15575 · NeurIPS · 2022

A Graph Is More Than Its Nodes: Towards Structured Uncertainty-Aware Learning on Graphs

Daniel Cremers, Hans Hao, Hsun Hsu, Yuesong Shen

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 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; 2 failures of our instrument; 1 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/2210.15575.json · badge: /api/ran/2210.15575.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
hans66hsu/structured_uncertainty_metrics canonical (classifier label)
classifier basis: classifier_2025
5 ran · 1 no recorded run
FunctionStatusWhere it lives
ECE Ran
ran on a synthesized input · constructed an object; no result was computed
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("a161c5cfbb1c7ed2")
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 an instance of its own class, ECE: the run constructed the object and did not call it, so no result was computed. 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
from torch import nn, Tensor, LongTensor, BoolTensor

torch.manual_seed(0)

# Create small sample tensors for testing
# confs: confidence scores (batch_size,)
# corrects: boolean tensor indicating correctness (batch_size,)
batch_size = 8
confs = torch.rand(batch_size)
corrects = torch.randint(0, 2, (batch_size,)).bool()

ARGS = {
    'bins': 4,
    'scheme': 'equal_width',
    'norm': 1
}
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 ECE — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ECE 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
EdgewiseECE Ran
ran on a synthesized input · constructed an object; no result was computed
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("6115863ca2e00949")
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 an instance of its own class, EdgewiseECE: the run constructed the object and did not call it, so no result was computed. 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
from torch import nn, Tensor, LongTensor

torch.manual_seed(0)

# Create small test tensors
# 10 nodes, 4 classes for logits
num_nodes = 10
num_classes = 4
logits = torch.randn(num_nodes, num_classes)

# Ground truth labels for each node
gts = torch.randint(0, num_classes, (num_nodes,))

# Edge index: pairs of node indices
# Create 8 edges as (source_nodes, target_nodes)
edge_index = torch.LongTensor([
    [0, 1, 2, 3, 4, 5, 6, 7],  # source indices
    [1, 2, 3, 4, 5, 6, 7, 8]   # target indices
])

ARGS = {
    'edge_index': edge_index,
    'bins': 15,
    'scheme': 'equal_width',
    'norm': 1
}
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 EdgewiseECE — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a EdgewiseECE 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
Reliability Ran
ran on a synthesized input · constructed an object; no result was computed
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("c9216e1dab899fc5")
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 an instance of its own class, Reliability: the run constructed the object and did not call it, so no result was computed. 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
from torch import nn, Tensor, LongTensor, BoolTensor

torch.manual_seed(0)

# Create small tensors for Reliability NamedTuple
conf = torch.randn(2, 4)  # confidence scores
acc = torch.randn(2, 3)   # accuracy scores (different dim)
count = torch.randint(0, 10, (2, 5), dtype=torch.long)  # count tensor

ARGS = {
    "conf": conf,
    "acc": acc,
    "count": count,
}
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 Reliability — the run constructed an instance of its own class and did not call it; no shape 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
edge_conf Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("70b486d8c4c102bb")
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 (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 nn, Tensor

torch.manual_seed(0)

# Create logits tensor: shape (num_nodes, num_classes)
# Using 6 nodes and 4 classes for variety
logits = torch.randn(6, 4)

# Create edge_index tensor: shape (2, num_edges)
# Each row contains node indices, forming edges between nodes
edge_index = torch.LongTensor([
    [0, 1, 2, 3],  # left node indices
    [1, 2, 3, 4]   # right node indices
])

ARGS = {
    "logits": logits,
    "edge_index": 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 1, dtype family float, shape (4,) — shape (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
partial_sums Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("dd7dae932e8116b6")
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 (2,) 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 nn, Tensor, LongTensor, BoolTensor

torch.manual_seed(0)

# t is a 1D tensor that will be summed according to groups
# The function treats t as elements that get grouped based on lens
t = torch.randn(5, dtype=torch.float32)

# lens is a LongTensor specifying the size of each group
# For example, if lens = [2, 3], it means first 2 elements go to group 0,
# next 3 elements go to group 1
# Sum of lens must equal t.size(0)
lens = torch.LongTensor([2, 3])

ARGS = {
    "t": t,
    "lens": lens,
}
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 (2,) — shape (2,) 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
EdgewiseBase Not yet run
unverified (harvested, no recorded run)
hans66hsu/structured_uncertainty_metrics/src/calibloss.py
code served (permissive licence) · get_code("351c6228b3e903a5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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 2026-11-22.

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 graph neural networks (GNNs) that tackle node classification on graphs tend to only focus on nodewise scores and are solely evaluated by nodewise metrics. This limits uncertainty estimation on graphs since nodewise marginals do not fully characterize the joint distribution given the graph structure. In this work, we propose novel edgewise metrics, namely the edgewise expected calibration error (ECE) and the agree/disagree ECEs, which provide criteria for uncertainty estimation on graphs beyond the nodewise setting. Our experiments demonstrate that the proposed edgewise metrics can complement the nodewise results and yield additional insights. Moreover, we show that GNN models which consider the structured prediction problem on graphs tend to have better uncertainty estimations, which illustrates the benefit of going beyond the nodewise setting. 2 * Equal contribution 2 Source code available at https://github.com/hans66hsu/structured_uncertainty_metrics

For agents

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

get_harvested_code_for_paper("2210.15575")
get_code_for_paper("2210.15575")
have("2210.15575")

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

curl https://syntology.ai/api/ran/2210.15575.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.15575.svg)](https://syntology.ai/paper/2210.15575)

Connect an agent — have() is free.