SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.05293 · 2023

Equivariant Graph Neural Networks for Charged Particle Tracking

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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, 2 violated it and 7 have no contract verdict; 1 failure of our instrument; 0 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/2304.05293.json · badge: /api/ran/2304.05293.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
ameya1101/equivariant-tracking canonical (classifier label)
classifier basis: census
10 ran · 0 no recorded run
FunctionStatusWhere it lives
R Ran
ran on a synthesized input
ameya1101/equivariant-tracking/equivariance_test-IN.py
code served (permissive licence) · get_code("a115a9638ba3c906")
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 (3, 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 torch

torch.manual_seed(0)

# The function R takes a single parameter: theta
# theta is used in np.cos(theta) and np.sin(theta)
# so it should be a numeric value (int, float, or scalar tensor)

theta = 0.5  # a simple float angle in radians

ARGS = {
    "theta": theta
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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 (3, 3) — shape (3, 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
build_dist_matrix Ran
ran on a synthesized input
ameya1101/equivariant-tracking/graph-construction/build_pre-clustering.py
code served (permissive licence) · get_code("33e8e7f4deba49f7")
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 (4, 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 numpy as np

# Analyze the function:
# - Takes X as input
# - X.shape[0] gets number of rows
# - X[i][0], X[i][1], X[i][2] are accessed (column indices 0, 1, 2)
# - X[i][2] is compared with == (appears to be a categorical/integer value)
# - X[i][0] and X[i][1] are used in arithmetic operations (eta and phi coordinates)
# - Returns a 2D distance matrix

# Create a small example dataset with 4 data points, each with at least 3 columns
# Columns represent: eta (float), phi (float), cluster_id (int)
X = np.array([
    [0.5, 1.0, 0],      # point 0, cluster 0
    [0.6, 1.1, 0],      # point 1, cluster 0
    [2.0, 3.5, 1],      # point 2, cluster 1
    [2.1, 3.6, 1],      # point 3, cluster 1
], dtype=np.float64)

ARGS = {
    "X": X
}
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 (4, 4) — shape (4, 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
calc_dphi Ran
ran, violated the contract we drafted
ameya1101/equivariant-tracking/graph-construction/build_data-driven.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from HEPTrkX/heptrkx-gnn-tracking, is pointer only (licence: NONE) · get_code("a32159421867451b")
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 HEPTrkX/heptrkx-gnn-tracking, 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.

calc_eta Ran
ran, violated the contract we drafted
ameya1101/equivariant-tracking/graph-construction/build_data-driven.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from HEPTrkX/heptrkx-gnn-tracking, is pointer only (licence: NONE) · get_code("5b89cdd0ef81e3dc")
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 HEPTrkX/heptrkx-gnn-tracking, 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.

cluster_hits Ran
ran on a synthesized input
ameya1101/equivariant-tracking/graph-construction/build_pre-clustering.py
code served (permissive licence) · get_code("4aa3702fe645d51f")
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 1, dtype family int, 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 numpy as np
from sklearn.cluster import DBSCAN

# Create a small distance matrix for clustering
# DBSCAN with metric='precomputed' expects a square distance matrix
# Create a 4x4 distance matrix (small but valid)
dist_matrix = np.array([
    [0.0, 0.02, 0.08, 0.3],
    [0.02, 0.0, 0.06, 0.25],
    [0.08, 0.06, 0.0, 0.35],
    [0.3, 0.25, 0.35, 0.0]
], dtype=np.float64)

ARGS = {
    'dist_matrix': dist_matrix,
    'eps': 0.05,
    'k': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, 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
euclidean_feats Ran
ran on a synthesized input
ameya1101/equivariant-tracking/utils.py
code served (permissive licence) · get_code("6d10099466848eda")
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
from typing import List, Union

torch.manual_seed(0)

# edge_index: Tensor with shape (2, num_edges)
# Contains source and target node indices
edge_index = torch.tensor([[0, 1, 0, 1], [1, 0, 1, 0]], dtype=torch.long)

# x: Tensor with shape (num_nodes, feature_dim)
# Node feature vectors
x = torch.randn(2, 4)

# s: Union[Tensor, None]
# Optional node features with shape (num_nodes, s_dim)
s = torch.randn(2, 3)

ARGS = {
    "edge_index": edge_index,
    "x": x,
    "s": s,
}
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
select_segments Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ameya1101/equivariant-tracking/graph-construction/build_geometric.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from GageDeZoort/interaction_network_paper, is pointer only (licence: NONE) · get_code("56512b2431599289")
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 GageDeZoort/interaction_network_paper, 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.

select_segments Ran
ran on a synthesized input
ameya1101/equivariant-tracking/graph-construction/build_data-driven.py
code served (permissive licence) · get_code("4365977a1a9f6716")
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 pandas as pd

# Create small sample dataframes for hits1 and hits2
np.random.seed(0)

# Create hits1 dataframe
hits1_data = {
    'evtid': [0, 0, 1, 1],
    'r': [10.5, 12.3, 11.1, 13.2],
    'phi': [0.5, 1.2, -2.8, 0.1],
    'z': [5.0, 15.0, -10.0, 20.0],
    'module_id': [0, 1, 0, 1]
}
hits1 = pd.DataFrame(hits1_data)

# Create hits2 dataframe
hits2_data = {
    'evtid': [0, 0, 1, 1],
    'r': [20.5, 22.3, 21.1, 23.2],
    'phi': [0.8, 1.5, -2.5, 0.4],
    'z': [25.0, 35.0, 10.0, 40.0],
    'module_id': [2, 3, 2, 3]
}
hits2 = pd.DataFrame(hits2_data)

# Create a module_map as a 2D numpy array
# This should be indexable by [mid1, mid2]
module_map = np.ones((4, 4), dtype=bool)

# Set some entries to False to test filtering
module_map[1, 3] = False

ARGS = {
    'hits1': hits1,
    'hits2': hits2,
    'phi_slope_max': 0.1,
    'z0_max': 50.0,
    'layer1': 0,
    'layer2': 4,
    'remove_intersecting_edges': False,
    'module_map': module_map
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
unsorted_segment_mean Ran
ran on a synthesized input
ameya1101/equivariant-tracking/utils.py
code served (permissive licence) · get_code("22dd60ad2b11a4bd")
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 (3, 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 Tensor

torch.manual_seed(0)

# Analyze the function:
# - data: Tensor with shape (batch_size, feature_dim) - used with .size(1), .new_zeros(), index_add_()
# - segment_ids: Tensor with shape (batch_size,) - used with index_add_()
# - num_segments: int - used as first dim of result tensor

# Create small test tensors
batch_size = 2
feature_dim = 4
num_segments = 3

# data: shape (batch_size, feature_dim)
data = torch.randn(batch_size, feature_dim)

# segment_ids: shape (batch_size,) with values in range [0, num_segments)
segment_ids = torch.tensor([0, 2], dtype=torch.long)

# num_segments: int indicating number of segments
num_segments_val = 3

ARGS = {
    "data": data,
    "segment_ids": segment_ids,
    "num_segments": num_segments_val
}
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 (3, 4) — shape (3, 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
unsorted_segment_sum Ran
ran on a synthesized input
ameya1101/equivariant-tracking/utils.py
code served (permissive licence) · get_code("483d183f742bc940")
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 (3, 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)

# Construct arguments for unsorted_segment_sum
# The function expects:
# - data: Tensor of shape (batch_size, feature_dim)
# - segment_ids: Tensor of shape (batch_size,) with segment indices
# - num_segments: int, the number of unique segments

batch_size = 2
feature_dim = 4
num_segments = 3

# Create data tensor: (2, 4)
data = torch.randn(batch_size, feature_dim)

# Create segment_ids tensor: (2,) with values in range [0, num_segments)
segment_ids = torch.tensor([0, 2], dtype=torch.long)

ARGS = {
    "data": data,
    "segment_ids": segment_ids,
    "num_segments": num_segments,
}
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 (3, 4) — shape (3, 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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 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, 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
calc_dphi 2 other papers
2 dated earlier
1810.06111 (2018-10)
Novel deep learning methods for track reconstruction
2103.16701 (2021-03)
Charged particle tracking via edge-classifying interaction networks
calc_eta 2 other papers
2 dated earlier
1810.06111 (2018-10)
Novel deep learning methods for track reconstruction
2103.16701 (2021-03)
Charged particle tracking via edge-classifying interaction networks
select_segments 1 other paper
1 dated earlier
2103.16701 (2021-03)
Charged particle tracking via edge-classifying interaction networks

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

Graph neural networks (GNNs) have gained traction in high-energy physics (HEP) for their potential to improve accuracy and scalability. However, their resource-intensive nature and complex operations have motivated the development of symmetry-equivariant architectures. In this work, we introduce EuclidNet, a novel symmetry-equivariant GNN for charged particle tracking. EuclidNet leverages the graph representation of collision events and enforces rotational symmetry with respect to the detector's beamline axis, leading to a more efficient model. We benchmark EuclidNet against the state-of-the-art Interaction Network on the TrackML dataset, which simulates high-pileup conditions expected at the High-Luminosity Large Hadron Collider (HL-LHC). Our results show that EuclidNet achieves near-state-of-the-art performance at small model scales (<1000 parameters), outperforming the non-equivariant benchmarks. This study paves the way for future investigations into more resource-efficient GNN models for particle tracking in HEP experiments.

For agents

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

get_harvested_code_for_paper("2304.05293")
get_code_for_paper("2304.05293")
have("2304.05293")

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

curl https://syntology.ai/api/ran/2304.05293.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.05293.svg)](https://syntology.ai/paper/2304.05293)

Connect an agent — have() is free.