SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2011.00186 · 2020

Self-supervised Representation Learning for Evolutionary Neural Architecture Search

arXiv · PDF · Open in the Atlas

Code that ran

10 of 13 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 10 have no contract verdict; 0 failures of our instrument; 3 no recorded run), measured 2026-11-21; 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/2011.00186.json · badge: /api/ran/2011.00186.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
auroua/SSNENAS canonical (classifier label)
classifier basis: census
10 ran · 3 no recorded run
FunctionStatusWhere it lives
acq_fn Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/algos_nas/acquisition_functions.py
code served (permissive licence) · get_code("8e50b14abde79c1b")
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
import torch

torch.manual_seed(0)

# For explore_type='ts' and 'its': predictions should be a numpy array
# with shape (num_samples, num_features)
predictions_numpy = np.random.randn(5, 4)

# For explore_type='its_vae': predictions should be a torch tensor
predictions_torch = torch.randn(5, 4)

ARGS = {
    'predictions': predictions_numpy,
    'explore_type': 'its',
    'reverse': False
}
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 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
algo_params_close_domain Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params_nas.py
code served (permissive licence) · get_code("96ee1ba824202f2b")
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 sys

# Define the ARGS dictionary with valid arguments for algo_params_close_domain
ARGS = {
    'param_str': 'nasbench_101',
    'search_budget': 100,
    'dataname': 'cifar10-valid'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
algo_params_open_domain Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params_nas.py
code served (permissive licence) · get_code("c38b7508cdb1f030")
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 numpy as np

torch.manual_seed(0)

ARGS = {
    'param_str': 'gin_predictor'
}
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 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
compute_best_test_losses Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/algos_nas/algo_compare.py
code served (permissive licence) · get_code("cb94f7ba1f7afae8")
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 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
import numpy as np

torch.manual_seed(0)

# The function expects `data` to be a list of items where each item is indexable
# and has at least indices 4 (val error) and 5 (test error)
# We'll create a list of tuples representing architecture evaluation records
# Format: (arch_id, param_count, train_error, val_error, val_error, test_error)

data = [
    (0, 100, 0.5, 0.45, 0.45, 0.47),      # arch 0: val_error=0.45, test_error=0.47
    (1, 150, 0.4, 0.35, 0.35, 0.38),      # arch 1: val_error=0.35, test_error=0.38
    (2, 120, 0.48, 0.42, 0.42, 0.44),     # arch 2: val_error=0.42, test_error=0.44
    (3, 180, 0.3, 0.28, 0.28, 0.31),      # arch 3: val_error=0.28, test_error=0.31
    (4, 110, 0.55, 0.50, 0.50, 0.52),     # arch 4: val_error=0.50, test_error=0.52
    (5, 160, 0.32, 0.29, 0.29, 0.32),     # arch 5: val_error=0.29, test_error=0.32
    (6, 140, 0.38, 0.33, 0.33, 0.36),     # arch 6: val_error=0.33, test_error=0.36
    (7, 170, 0.31, 0.27, 0.27, 0.30),     # arch 7: val_error=0.27, test_error=0.30
    (8, 130, 0.44, 0.40, 0.40, 0.42),     # arch 8: val_error=0.40, test_error=0.42
    (9, 190, 0.29, 0.26, 0.26, 0.29),     # arch 9: val_error=0.26, test_error=0.29
]

k = 2
total_queries = 10

ARGS = {
    "data": data,
    "k": k,
    "total_queries": total_queries,
}
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 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
evolution_search_nasbench_101 Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/algos_nas/evolution.py
code served (permissive licence) · get_code("7f5760af9d9e1fdb")
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. 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 random
import numpy as np
from logging import Logger
import logging

# Create a minimal logger
logger = logging.getLogger('test_logger')
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
logger.addHandler(handler)

# Create a minimal SearchSpace class that satisfies the function's needs
class MinimalSearchSpace:
    def __init__(self):
        self.call_count = 0
    
    def generate_random_dataset(self, num=10, allow_isomorphisms=False, deterministic_loss=True):
        """Generate random dataset with architecture data and validation loss"""
        dataset = []
        for i in range(num):
            # Each item: (id, matrix, ops, ..., val_loss)
            matrix = np.random.randint(0, 2, size=(5, 5))
            ops = ['conv3x3', 'conv1x1', 'maxpool3x3']
            val_loss = np.random.random()
            dataset.append((i, matrix, ops, None, val_loss))
        return dataset
    
    def mutate_arch(self, arch, mutation_rate):
        """Mutate an architecture and return new data tuple"""
        matrix = arch['matrix'].copy()
        ops = arch['ops'].copy()
        
        # Simple mutation: flip a random bit
        if np.random.random() < mutation_rate:
            i, j = np.random.randint(0, matrix.shape[0], 2)
            matrix[i, j] = 1 - matrix[i, j]
        
        val_loss = np.random.random()
        # Return tuple with same structure as original dataset items
        return (len([]), matrix, ops, None, val_loss)

search_space = MinimalSearchSpace()

ARGS = {
    'search_space': search_space,
    'num_init': 10,
    'k': 10,
    'population_size': 30,
    'total_queries': 100,
    'tournament_size': 10,
    'mutation_rate': 1.0,
    'allow_isomorphisms': False,
    'deterministic': True,
    'verbose': 1,
    'logger': logger,
}
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 list — a list 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
meta_neuralnet_params Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params_nas.py
code served (permissive licence) · get_code("fd6439b6def33330")
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 sys

# The function takes a single parameter: param_str
# It expects a string that is either 'nasbench_101' or 'nasbench_201'
# to return valid params, or any other string will cause sys.exit()

ARGS = {
    'param_str': 'nasbench_101'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
quick_sort_list Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/algos/algo_sort.py
code served (permissive licence) · get_code("383278a1ef55d3cd")
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. 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 copy

def _compare_lists(list1, list2):
    for l1, l2 in zip(list1[1], list2[1]):
        if l1 == l2:
            continue
        elif l1 < l2:
            return True
        else:
            return False

def _partition_list(nums, low, high):
    pivot = nums[(low + high) // 2]
    i = low - 1
    j = high + 1
    while True:
        i += 1
        while _compare_lists(nums[i], pivot):
            i += 1
        j -= 1
        while _compare_lists(pivot, nums[j]):
            j -= 1
        if i >= j:
            return j
        nums[i], nums[j] = (nums[j], nums[i])

def quick_sort_list(nums):
    datas = copy.deepcopy(nums)

    def _quick_sort(items, low, high):
        if low < high:
            split_index = _partition_list(items, low, high)
            _quick_sort(items, low, split_index)
            _quick_sort(items, split_index + 1, high)
    _quick_sort(datas, 0, len(datas) - 1)
    return datas

# Construct test data
# The function expects nums to be a list of elements where each element has [1] indexable
# and those elements are comparable
# Based on _compare_lists, each item in nums should be a sequence where index [1] 
# contains a comparable sequence

nums = [
    (0, [5, 3, 1]),
    (1, [2, 8, 4]),
    (2, [1, 1, 7]),
    (3, [3, 2, 9])
]

ARGS = {
    "nums": nums
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list 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
supervised_ss_rl_params Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params.py
code served (permissive licence) · get_code("33da0703810ce533")
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 sys

def supervised_ss_rl_params(param_str):
    if param_str == 'nasbench_101':
        params = {'lr': 0.005, 'batch_size': 64, 'input_dim': 6}
    elif param_str == 'nasbench_201':
        params = {'lr': 0.001, 'batch_size': 64, 'input_dim': 8}
    elif param_str == 'darts':
        params = {'lr': 0.005, 'batch_size': 64, 'input_dim': 11}
    else:
        print('invalid params')
        sys.exit()
    return params

ARGS = {
    'param_str': 'nasbench_101'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
unsupervised_ss_rl_params Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params.py
code served (permissive licence) · get_code("4aa9f462c2fbfbcb")
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 sys

def unsupervised_ss_rl_params(param_str):
    if param_str == 'nasbench_101':
        params = {'lr': 0.0005, 'batch_size': 64, 'input_dim': 6}
    elif param_str == 'nasbench_201':
        params = {'lr': 0.001, 'batch_size': 64, 'input_dim': 8}
    elif param_str == 'darts':
        params = {'lr': 0.0005, 'batch_size': 64, 'input_dim': 11}
    else:
        print('invalid params')
        sys.exit()
    return params

ARGS = {
    'param_str': 'nasbench_101'
}
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
unsupervised_ss_rl_pre_train Ran
ran on a synthesized input
auroua/SSNENAS/nas_lib/params.py
code served (permissive licence) · get_code("ac8329f01f864a9e")
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 sys

torch.manual_seed(0)

ARGS = {
    'param_str': 'nasbench_101'
}
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
build_open_algos Not yet run
unverified (harvested, no recorded run)
auroua/SSNENAS/nas_lib/algos_darts/build_open_algos.py
code served (permissive licence) · get_code("b86e1c4f11405bbc")
How it ran

no invocation recorded.

run_nas_algos_nasbench_101 Not yet run
unverified (harvested, no recorded run)
auroua/SSNENAS/nas_lib/algos_nas/algo_compare.py
code served (permissive licence) · get_code("d9cc4398d1baf37a")
How it ran

no invocation recorded.

run_nas_algos_nasbench_201 Not yet run
unverified (harvested, no recorded run)
auroua/SSNENAS/nas_lib/algos_nas/algo_compare.py
code served (permissive licence) · get_code("d6d169f03ec8c6ea")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 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 13 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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Recently proposed neural architecture search (NAS) algorithms adopt neural predictors to accelerate the architecture search. The capability of neural predictors to accurately predict the performance metrics of neural architecture is critical to NAS, and the acquisition of training datasets for neural predictors is time-consuming. How to obtain a neural predictor with high prediction accuracy using a small amount of training data is a central problem to neural predictor-based NAS. Here, we firstly design a new architecture encoding scheme that overcomes the drawbacks of existing vector-based architecture encoding schemes to calculate the graph edit distance of neural architectures. To enhance the predictive performance of neural predictors, we devise two self-supervised learning methods from different perspectives to pre-train the architecture embedding part of neural predictors to generate a meaningful representation of neural architectures. The first one is to train a carefully designed two branch graph neural network model to predict the graph edit distance of two input neural architectures. The second method is inspired by the prevalently contrastive learning, and we present a new contrastive learning algorithm that utilizes a central feature vector as a proxy to contrast positive pairs against negative pairs. Experimental results illustrate that the pre-trained neural predictors can achieve comparable or superior performance compared with their supervised counterparts with several times less training samples. We achieve state-of-the-art performance on the NASBench-101 and NASBench201 benchmarks when integrating the pre-trained neural predictors with an evolutionary NAS algorithm.

For agents

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

get_harvested_code_for_paper("2011.00186")
get_code_for_paper("2011.00186")
have("2011.00186")

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

curl https://syntology.ai/api/ran/2011.00186.json

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

[![Syntology run record](https://syntology.ai/api/ran/2011.00186.svg)](https://syntology.ai/paper/2011.00186)

Connect an agent — have() is free.