SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2201.09332 · 2022

How Expressive are Transformers in Spectral Domain for Graphs?

arXiv · PDF · Open in the Atlas

Code that ran

4 of 6 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 2 have no contract verdict; 2 failures of our instrument; 2 no recorded run), measured 2026-11-22; 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/2201.09332.json · badge: /api/ran/2201.09332.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
ansonb/FeTA_TMLR canonical (classifier label)
classifier basis not recorded on the link
4 ran · 2 no recorded run
FunctionStatusWhere it lives
cheb_msg Ran
ran on a synthesized input
ansonb/FeTA_TMLR/LPE/layers/gat_feta_layer.py
code served (permissive licence) · get_code("1b26e6ea6d2db1d1")
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 function. 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
import torch.nn as nn

torch.manual_seed(0)

# The function cheb_msg expects three string parameters that represent field names.
# It returns a function that operates on an edges object (like DGL edges) that has:
# - edges.src[src_field]: source node feature
# - edges._edge_data[dst_field]: edge data
# - The returned dict uses out_field as a key

# Create mock edge object for understanding the structure
class MockEdges:
    def __init__(self):
        self.src = {}
        self._edge_data = {}

# Set up small tensors for the fields
src_tensor = torch.randn(2, 4)  # batch_size=2, feature_dim=4
edge_tensor = torch.randn(2, 4)  # batch_size=2, feature_dim=4

# Create a mock edges object with the required fields
mock_edges = MockEdges()
mock_edges.src = {"node_features": src_tensor}
mock_edges._edge_data = {"edge_features": edge_tensor}

# Define the arguments
ARGS = {
    "src_field": "node_features",
    "dst_field": "edge_features",
    "out_field": "output"
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
gcn_msg Ran
ran on a synthesized input
ansonb/FeTA_TMLR/LPE/layers/gat_feta_layer.py
code served (permissive licence) · get_code("85162c58be024fa2")
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 function. 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)

# The function gcn_msg returns a closure that operates on edges
# The closure accesses:
# - edges.src[src_field]: source node features
# - edges._edge_data[dst_field]: edge data
# - Returns a dict with out_field as key

# We need string parameters for the field names
src_field = "node_features"
dst_field = "edge_weight"
out_field = "message"

ARGS = {
    "src_field": src_field,
    "dst_field": dst_field,
    "out_field": out_field,
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
imp_exp_attn Ran
ran; our contract draft was wrong, not the code
ansonb/FeTA_TMLR/LPE/layers/GT_layer_bkp.py
code served (permissive licence) · get_code("700de849477f93ce")
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 function. 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
import torch.nn as nn

torch.manual_seed(0)

# The function creates a closure that accesses edges.data[implicit_attn] and edges.data[explicit_edge]
# These are string keys used to access edge data in a DGL-like edge batch object
# We need to construct a mock edge batch object that has a .data dict

class MockEdgeBatch:
    def __init__(self):
        # Create mock data with tensors
        self.data = {
            'attn': torch.randn(2, 4),
            'edge_feat': torch.randn(2, 4)
        }

# Create the arguments
implicit_attn = 'attn'
explicit_edge = 'edge_feat'

ARGS = {
    'implicit_attn': implicit_attn,
    'explicit_edge': explicit_edge
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
scaling Ran
ran; our contract draft was wrong, not the code
ansonb/FeTA_TMLR/LPE/layers/GT_layer_bkp.py
code served (permissive licence) · get_code("a3fb87a259b15eae")
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 function. 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
import torch.nn as nn

torch.manual_seed(0)

# The function `scaling` returns a function that expects an object with a `data` attribute
# The `data` attribute should be a dict-like object with the field specified
# We need to create an object that mimics what DGL uses (an edge batch object)

class MockEdges:
    def __init__(self):
        self.data = {
            'features': torch.tensor([[1.0, 2.0, 3.0, 4.0],
                                      [5.0, 6.0, 7.0, 8.0]])
        }

ARGS = {
    'field': 'features',
    'scale_constant': 2.0
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
send_source Not yet run
unverified (harvested, no recorded run)
ansonb/FeTA_TMLR/LPE/layers/gat_feta_layer.py
code served (permissive licence) · get_code("7e094af0e4a3779e")
How it ran

no invocation recorded.

src_dot_dst Not yet run
unverified (harvested, no recorded run)
ansonb/FeTA_TMLR/LPE/layers/GT_layer_bkp.py
code served (permissive licence) · get_code("9408c2ca6e799410")
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, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 3 later) and 3 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
scaling 6 other papers
1 dated earlier, 5 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2308.08173 (2023-08)
Expressivity of Graph Neural Networks Through the Lens of Adversarial Robustness
2308.09517 (2023-08)
Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-based Similarity
and 3 more not named here
src_dot_dst 6 other papers
1 dated earlier, 5 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2308.08173 (2023-08)
Expressivity of Graph Neural Networks Through the Lens of Adversarial Robustness
2308.09517 (2023-08)
Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-based Similarity
and 3 more not named here
imp_exp_attn 5 other papers
1 dated earlier, 4 later
2110.07875 (2021-10)
Graph Neural Networks with Learnable Structural and Positional Representations
2308.08173 (2023-08)
Expressivity of Graph Neural Networks Through the Lens of Adversarial Robustness
2308.09517 (2023-08)
Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-based Similarity
and 2 more not named here

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

The recent works proposing transformer-based models for graphs have proven the inadequacy of Vanilla Transformer for graph representation learning. To understand this inadequacy, there is a need to investigate if spectral analysis of the transformer will reveal insights into its expressive power. Similar studies already established that spectral analysis of Graph neural networks (GNNs) provides extra perspectives on their expressiveness. In this work, we systematically study and establish the link between the spatial and spectral domain in the realm of the transformer. We further provide a theoretical analysis and prove that the spatial attention mechanism in the transformer cannot effectively capture the desired frequency response, thus, inherently limiting its expressiveness in spectral space. Therefore, we propose FeTA, a framework that aims to perform attention over the entire graph spectrum (i.e., actual frequency components of the graphs) analogous to the attention in spatial space. Empirical results suggest that FeTA provides homogeneous performance gain against vanilla transformer across all tasks on standard benchmarks and can easily be extended to GNN-based models with low-pass characteristics (e.g., GAT).

For agents

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

get_harvested_code_for_paper("2201.09332")
get_code_for_paper("2201.09332")
have("2201.09332")

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

curl https://syntology.ai/api/ran/2201.09332.json

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

[![Syntology run record](https://syntology.ai/api/ran/2201.09332.svg)](https://syntology.ai/paper/2201.09332)

Connect an agent — have() is free.