SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.06936 · NeurIPS · 2020

Does Unsupervised Architecture Representation Learning Help Neural Architecture Search?

Shen Yan, Yu Zheng, Wei Ao, Xiao Zeng, Mi Zhang

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 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 7 have no contract verdict; 0 failures of our instrument; 2 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/2006.06936.json · badge: /api/ran/2006.06936.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
MSU-MLSys-Lab/arch2vec canonical (classifier label)
classifier basis: census
7 ran · 2 no recorded run
FunctionStatusWhere it lives
edge_match Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/plot_scripts/try_networkx.py
code served (permissive licence) · get_code("720e98759d5548ad")
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 bool, 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 numpy as np

torch.manual_seed(0)

# The function edge_match takes two parameters e1 and e2
# With no body implementation, we just need to provide valid arguments
# Based on the function signature, e1 and e2 are likely edge representations

# Create small tensor representations for edges
e1 = torch.randn(2, 4)  # batch of 2 edges, 4 features each
e2 = torch.randn(2, 4)  # batch of 2 edges, 4 features each

ARGS = {
    "e1": e1,
    "e2": e2,
}
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 bool, 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
gen_graph Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/plot_scripts/try_networkx.py
code served (permissive licence) · get_code("bcb71528933a9cb0")
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 DiGraph. 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 numpy as np
import networkx as nx

torch.manual_seed(0)

# Create a small adjacency matrix
# Shape: (4, 4) - 4 nodes
# Upper triangular to represent DAG structure
adj = np.array([
    [0, 1, 1, 0],
    [0, 0, 1, 1],
    [0, 0, 0, 1],
    [0, 0, 0, 0]
], dtype=np.int32)

# Create ops list with 4 elements (one for each node)
ops = ['conv', 'relu', 'pool', 'fc']

ARGS = {
    'adj': adj,
    'ops': ops
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DiGraph — a DiGraph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DiGraph output the recorded ndim 1 and dtype family int 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
load_arch2vec Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/search_methods/dngo.py
code served (permissive licence) · get_code("5da0dc870e7a3deb")
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
import tempfile
import os

torch.manual_seed(0)

# Create a temporary file to store the embedding data
temp_dir = tempfile.gettempdir()
embedding_path = os.path.join(temp_dir, 'test_arch2vec_embedding.pt')

# Create a small embedding dictionary with the required structure
embedding = []
for i in range(2):
    embedding.append({
        'feature': torch.randn(4),
        'valid_accuracy': 0.85 + i * 0.05,
        'test_accuracy': 0.82 + i * 0.06,
        'time': 100.0 + i * 10.0
    })

# Save the embedding to the temporary file
torch.save(embedding, embedding_path)

ARGS = {
    'embedding_path': embedding_path
}
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
load_arch2vec Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/search_methods/dngo_darts.py
code served (permissive licence) · get_code("8579b264cb1a5015")
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
import tempfile
import os

torch.manual_seed(0)

# Create a temporary directory and file for the embedding
temp_dir = tempfile.mkdtemp()
embedding_path = os.path.join(temp_dir, "embedding.pt")

# Create sample embedding data matching the function's expectations
# The function expects a list/dict where each element has 'feature' and 'genotype' keys
embedding_data = []
for i in range(2):
    embedding_data.append({
        'feature': torch.randn(4),
        'genotype': f'genotype_{i}'
    })

# Save the embedding data
torch.save(embedding_data, embedding_path)

ARGS = {
    'embedding_path': embedding_path
}
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
node_match Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/plot_scripts/try_networkx.py
code served (permissive licence) · get_code("682ad0b128a9b119")
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 bool, 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)

# The function node_match takes two nodes (dicts) and compares their 'op' key
# We need to construct two dicts with an 'op' key

n1 = {'op': 'conv2d', 'other_field': 'value1'}
n2 = {'op': 'conv2d', 'other_field': 'value2'}

ARGS = {
    'n1': n1,
    'n2': n2
}
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 bool, 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
process Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/models/pretraining_darts.py
code served (permissive licence) · get_code("8b92c5553aa736fc")
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 list, ndim 2, dtype family int; 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 torch
import numpy as np

# Construct valid arguments for the process function
# The function iterates over geno and converts each item to a tuple
# geno should be an iterable of list-like objects

ARGS = {
    "geno": [
        [1, 2, 3, 4],
        [5, 6, 7, 8],
    ]
}
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 list, ndim 2, dtype family int — a list 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
transform_operations Ran
ran on a synthesized input
MSU-MLSys-Lab/arch2vec/models/pretraining_nasbench101.py
code served (permissive licence) · get_code("d59c36746025bd19")
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 list, 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 torch

torch.manual_seed(0)

# The function expects max_idx to be an iterable of tensor elements
# Each element should be indexable with .item() to get an integer
# The integers should be keys in transform_dict (0-4)

max_idx = [torch.tensor(0), torch.tensor(1), torch.tensor(2), torch.tensor(3), torch.tensor(4)]

ARGS = {
    'max_idx': max_idx
}
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 list, ndim 1, dtype family float — a list 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
get_init_samples Not yet run
unverified (harvested, no recorded run)
MSU-MLSys-Lab/arch2vec/search_methods/dngo.py
code served (permissive licence) · get_code("c95b21a97a430d4b")
How it ran

no invocation recorded.

propose_location Not yet run
unverified (harvested, no recorded run)
MSU-MLSys-Lab/arch2vec/search_methods/dngo.py
code served (permissive licence) · get_code("0c8092d946ec882a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 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 9 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

Existing Neural Architecture Search (NAS) methods either encode neural architectures using discrete encodings that do not scale well, or adopt supervised learning-based methods to jointly learn architecture representations and optimize architecture search on such representations which incurs search bias. Despite the widespread use, architecture representations learned in NAS are still poorly understood. We observe that the structural properties of neural architectures are hard to preserve in the latent space if architecture representation learning and search are coupled, resulting in less effective search performance. In this work, we find empirically that pre-training architecture representations using only neural architectures without their accuracies as labels improves the downstream architecture search efficiency. To explain this finding, we visualize how unsupervised architecture representation learning better encourages neural architectures with similar connections and operators to cluster together. This helps map neural architectures with similar performance to the same regions in the latent space and makes the transition of architectures in the latent space relatively smooth, which considerably benefits diverse downstream search strategies. 34th Conference on Neural Information Processing Systems (NeurIPS 2020),

For agents

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

get_harvested_code_for_paper("2006.06936")
get_code_for_paper("2006.06936")
have("2006.06936")

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

curl https://syntology.ai/api/ran/2006.06936.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.06936.svg)](https://syntology.ai/paper/2006.06936)

Connect an agent — have() is free.