SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2311.05784 · 2023

Are "Hierarchical" Visual Representations Hierarchical?

arXiv · PDF · Open in the Atlas

Code that ran

7 of 8 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 0 have no contract verdict; 7 failures of our instrument; 1 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/2311.05784.json · badge: /api/ran/2311.05784.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
ethanlshen/hiernet canonical (classifier label)
classifier basis: census
7 ran · 1 no recorded run
FunctionStatusWhere it lives
cluster_mode Ran
ran; our contract draft was wrong, not the code
ethanlshen/hiernet/cluster_analysis.py
code served (permissive licence) · get_code("134dc398b96b14bb")
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 int, shape (3, 1) 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

# Construct valid arguments for cluster_mode function
pred_labels = np.array([0, 0, 1, 1, 2, 2])
true_labels = np.array([5, 5, 7, 8, 9, 9])

ARGS = {
    "pred_labels": pred_labels,
    "true_labels": true_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 int, shape (3, 1) — shape (3, 1) 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
convert_pca Ran
ran; our contract draft was wrong, not the code
ethanlshen/hiernet/embeddings.py
code served (permissive licence) · get_code("283da2a58b525a29")
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 dict. 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)

# Create a minimal ResNet-like model with fc layer
class MinimalResNet(nn.Module):
    def __init__(self):
        super().__init__()
        # fc.weight will be used in SVD: model.fc.weight.T
        # The function does SVD on (n_features, n_classes)
        # embed is (n_samples, 2048), so we need fc.weight to be (n_classes, 2048)
        self.fc = nn.Linear(2048, 10)
    
    def forward(self, x):
        return self.fc(x)

model = MinimalResNet()

# Create embed_dict with the expected structure
n_samples = 2
embed_dim = 2048
n_classes = 10

embed_dict = {
    'embed': torch.randn(n_samples, embed_dim),
    'labels': torch.tensor([0, 1]),
    'preds': torch.tensor([0, 1]),
    'subclass_labels': torch.tensor([0, 1])
}

# Parameters for convert_pca
dim = 4
device = 'cpu'

ARGS = {
    'model': model,
    'dim': dim,
    'device': device,
    'embed_dict': embed_dict
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
hhot Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ethanlshen/hiernet/distances.py
code served (permissive licence) · get_code("91b992f48a589147")
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 ot

# Create a small example with 2 samples, 2 true clusters, 2 predicted clusters
n_samples = 4

# True labels: 2 samples in cluster 0, 2 samples in cluster 1
true_labels = np.array([0, 0, 1, 1])

# Predicted labels: samples distributed differently across 2 predicted clusters
pred_labels = np.array([0, 1, 0, 1])

# Cost matrix (4x4) - pairwise costs between samples
# Use a simple symmetric cost matrix
cost = np.array([
    [0.0, 1.0, 2.0, 3.0],
    [1.0, 0.0, 1.5, 2.5],
    [2.0, 1.5, 0.0, 1.0],
    [3.0, 2.5, 1.0, 0.0]
])

ARGS = {
    "true_labels": true_labels,
    "pred_labels": pred_labels,
    "cost": cost
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, ot — 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
hhot_h2 Ran
ran; our contract draft was wrong, not the code
ethanlshen/hiernet/distances.py
code served (permissive licence) · get_code("db29085c79c982b3")
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 float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 ot

# Create a simple cluster cost matrix
# 4 true clusters (axis 0) x 3 predicted clusters (axis 1)
cluster_cost = np.array([
    [0.1, 0.5, 0.9],
    [0.2, 0.3, 0.8],
    [0.6, 0.2, 0.4],
    [0.7, 0.1, 0.3]
], dtype=np.float64)

# true_superclass_merge: 2 superclasses
# Superclass 0 contains subclasses 0, 1
# Superclass 1 contains subclasses 2, 3
true_superclass_merge = {
    0: {0: None, 1: None},
    1: {2: None, 3: None}
}

# pred_superclass_merge: 2 superclasses
# Superclass 0 contains subclasses 0
# Superclass 1 contains subclasses 1, 2
pred_superclass_merge = {
    0: {0: None},
    1: {1: None, 2: None}
}

ARGS = {
    "cluster_cost": cluster_cost,
    "true_superclass_merge": true_superclass_merge,
    "pred_superclass_merge": pred_superclass_merge
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, ot — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
image_distance Ran
ran; our contract draft was wrong, not the code
ethanlshen/hiernet/distances.py
code served (permissive licence) · get_code("6a5bc9309999a232")
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 two small image representation tensors
# Shape: (x_dim, y_dim) = (4, 3)
i1 = torch.randn(4, 3)
i2 = torch.randn(4, 3)

ARGS = {
    "i1": i1,
    "i2": i2,
}
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
metrics Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ethanlshen/hiernet/cluster_analysis.py
code served (permissive licence) · get_code("00dec92e5e70554d")
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, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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

# Create small sample data for clustering evaluation
# true_labels: ground truth cluster assignments
# pred_labels: predicted cluster assignments

true_labels = np.array([0, 0, 1, 1, 2, 2])
pred_labels = np.array([0, 0, 1, 1, 1, 2])

ARGS = {
    "true_labels": true_labels,
    "pred_labels": pred_labels,
}
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 tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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
purity_score Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ethanlshen/hiernet/cluster_analysis.py
code served (permissive licence) · get_code("9db7d81587e73349")
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 float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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

# Create small sample data for clustering evaluation
# true_labels: ground truth cluster assignments
# pred_labels: predicted cluster assignments

np.random.seed(0)

# Create 8 samples with ground truth labels (0, 1, or 2)
true_labels = np.array([0, 0, 1, 1, 2, 2, 0, 1])

# Create predicted labels that partially match ground truth
pred_labels = np.array([0, 0, 1, 0, 2, 2, 1, 1])

ARGS = {
    "true_labels": true_labels,
    "pred_labels": pred_labels,
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
embed_breeds Not yet run
unverified (harvested, no recorded run)
ethanlshen/hiernet/embeddings.py
code served (permissive licence) · get_code("df8d9bffa5bad410")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 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 8 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

Learned visual representations often capture large amounts of semantic information for accurate downstream applications. Human understanding of the world is fundamentally grounded in hierarchy. To mimic this and further improve representation capabilities, the community has explored "hierarchical" visual representations that aim at modeling the underlying hierarchy of the visual world. In this work, we set out to investigate if hierarchical visual representations truly capture the human perceived hierarchy better than standard learned representations. To this end, we create HierNet, a suite of 12 datasets spanning 3 kinds of hierarchy from the BREEDs subset of ImageNet. After extensive evaluation of Hyperbolic and Matryoshka Representations across training setups, we conclude that they do not capture hierarchy any better than the standard representations but can assist in other aspects like search efficiency and interpretability. Our benchmark and the datasets are open-sourced at https://github.com/ethanlshen/HierNet.

For agents

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

get_harvested_code_for_paper("2311.05784")
get_code_for_paper("2311.05784")
have("2311.05784")

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

curl https://syntology.ai/api/ran/2311.05784.json

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

[![Syntology run record](https://syntology.ai/api/ran/2311.05784.svg)](https://syntology.ai/paper/2311.05784)

Connect an agent — have() is free.