SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2311.06295 · 2023

Gradual Optimization Learning for Conformational Energy Minimization

arXiv · PDF · Open in the Atlas

Code that ran

9 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 7 failures 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/2311.06295.json · badge: /api/ran/2311.06295.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
airi-institute/golf canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
FunctionStatusWhere it lives
aggregate2string Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/evaluate_batch_dft.py
code served (permissive licence) · get_code("bb450a928edb1ffa")
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 math
import numpy as np

# The function processes a stats dictionary with various numeric values
# Some keys contain 'pct' and are filtered by pct thresholds
# All values are converted to numpy arrays and processed

ARGS = {
    'stats': {
        'loss': [0.5, 0.6, 0.7, 0.8],
        'accuracy': [0.85, 0.87, 0.89, 0.91],
        'pct_improvement': [5.2, 6.1, 7.3, 8.0],
        'pct_change': [2.5, 3.1, 4.2, 5.5],
    },
    'pct_min_threshold': 0.0,
    'pct_max_threshold': 100.0,
    'digits': 3,
}
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 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
calculate_dft_energy_tcp_client Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/env/dft.py
code served (permissive licence) · get_code("2e63244f4af8ee6b")
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 pickle
import socket
import struct
from datetime import datetime
from unittest.mock import Mock, MagicMock, patch
import sys

# Create a mock ASE Atoms object
class MockASEAtoms:
    def __init__(self):
        self.positions = [[0.0, 0.0, 0.0], [1.0, 1.0, 1.0]]
        self.numbers = [1, 1]
        self.cell = [[10.0, 0.0, 0.0], [0.0, 10.0, 0.0], [0.0, 0.0, 10.0]]
        self.pbc = [True, True, True]
    
    def todict(self):
        return {
            'positions': self.positions,
            'numbers': self.numbers,
            'cell': self.cell,
            'pbc': self.pbc
        }

# Create task tuple
conformation_id = 42
step = 10
ase_atoms = MockASEAtoms()
task = (conformation_id, step, ase_atoms)

# Create arguments dict
ARGS = {
    'task': task,
    'host': 'localhost',
    'port': 9999,
    'logging': False
}
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 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
get_lr_scheduler Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/GOLF/utils.py
code served (permissive licence) · get_code("7c532365402d1a64")
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 OneCycleLR. 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
import torch.optim as optim

torch.manual_seed(0)

# Create a simple model to optimize
model = nn.Linear(4, 2)

# Create an optimizer (required for all scheduler types)
optimizer = optim.SGD(model.parameters(), lr=0.1)

# Define ARGS mapping all parameters of get_lr_scheduler
ARGS = {
    'scheduler_type': 'OneCycleLR',
    'optimizer': optimizer,
    'initial_lr': 0.1,
    'final_div_factor': 10,
    'total_steps': 100,
    'last_epoch': -1,
    'gamma': 0.1
}
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 OneCycleLR — a OneCycleLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OneCycleLR 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
mean_std Ran
ran, honoured the contract we drafted
airi-institute/golf/read_evaluation_metrics.py
code served (permissive licence) · get_code("d7a78068f73dfc9f")
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 math
import numpy as np

# The function mean_std takes:
# - values: array-like that will be converted to np.ndarray
# - min_threshold: numeric value (default -math.inf)
# - max_threshold: numeric value (default math.inf)

# Construct test values
values = [1.5, 2.3, 3.1, 4.2, 5.8, 6.1, 7.0, 8.5]

# Set thresholds that will filter some values
min_threshold = 2.0
max_threshold = 7.5

ARGS = {
    "values": values,
    "min_threshold": min_threshold,
    "max_threshold": max_threshold,
}
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 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
outliers_fraction Ran
ran, honoured the contract we drafted
airi-institute/golf/read_evaluation_metrics.py
code served (permissive licence) · get_code("f2718c3c137d1590")
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 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 math
import numpy as np

# The function takes:
# - values: array-like (will be converted to numpy array)
# - min_threshold: numeric, default -inf
# - max_threshold: numeric, default +inf

# Create a simple array of values with some outliers
values = np.array([1.0, 2.0, 3.0, 4.0, 5.0, -10.0, 20.0])

ARGS = {
    "values": values,
    "min_threshold": 0.0,
    "max_threshold": 10.0
}
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 tuple, ndim 2, 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
read_metrics Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/evaluate_batch_dft.py
code served (permissive licence) · get_code("892727a56439fd79")
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 json
import tempfile
import os

# Create a temporary file with JSON lines (metrics)
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.jsonl')
temp_path = temp_file.name

# Write some sample metrics as JSON lines
metrics_data = [
    {"loss": 0.5, "accuracy": 0.95},
    {"loss": 0.3, "accuracy": 0.97},
    {"loss": 0.2, "accuracy": 0.98},
]

for metric in metrics_data:
    temp_file.write(json.dumps(metric) + '\n')

# Also add an empty line to test the empty line handling
temp_file.write('\n')

# Add one more metric
temp_file.write(json.dumps({"loss": 0.1, "accuracy": 0.99}) + '\n')

temp_file.close()

ARGS = {
    "path": temp_path
}
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
read_metrics Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/read_evaluation_metrics.py
code served (permissive licence) · get_code("9eae9feb6656ee86")
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 tempfile
import json
import os

# Create a temporary file with metrics data in JSONL format
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.jsonl')
temp_path = temp_file.name

# Write sample metrics data
sample_metrics = [
    {'conformation_id': 'conf_001', 'energy': 3.14, 'rmsd': 0.5},
    {'conformation_id': 'conf_002', 'energy': 2.71, 'rmsd': 0.3},
    {'conformation_id': 'conf_003', 'energy': 1.41, 'rmsd': 0.7},
]

for metric in sample_metrics:
    temp_file.write(json.dumps(metric) + '\n')

temp_file.close()

ARGS = {
    'path': temp_path
}
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
recv_one_message Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/env/dft.py
code served (permissive licence) · get_code("6061e72c09ebd282")
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 bytes. 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 socket
import struct

# Create a mock socket that behaves like a real socket for testing
class MockSocket:
    def __init__(self, data):
        self.data = data
        self.pos = 0
    
    def recv(self, count):
        if self.pos >= len(self.data):
            return b''
        chunk = self.data[self.pos:self.pos + count]
        self.pos += len(chunk)
        return chunk

# Construct test data: a 4-byte length header followed by message data
message_content = b"Hello, World!"
message_length = len(message_content)
length_header = struct.pack('!I', message_length)
test_data = length_header + message_content

# Create mock socket with test data
mock_sock = MockSocket(test_data)

ARGS = {
    'sock': mock_sock
}
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 bytes — a bytes output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a bytes 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
recvall Ran
ran; our contract draft was wrong, not the code
airi-institute/golf/env/dft.py
code served (permissive licence) · get_code("f33201b615cc0380")
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 bytes. 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 socket
from unittest.mock import Mock, MagicMock

# Create a mock socket object that behaves like a real socket
mock_sock = Mock(spec=socket.socket)

# Simulate receiving data in chunks
# First call returns 5 bytes, second call returns 5 bytes, third call returns 0 (connection closed)
mock_sock.recv.side_effect = [b'hello', b'world', b'']

ARGS = {
    'sock': mock_sock,
    'count': 10
}
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 bytes — a bytes output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a bytes 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

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

Molecular conformation optimization is crucial to computer-aided drug discovery and materials design. Traditional energy minimization techniques rely on iterative optimization methods that use molecular forces calculated by a physical simulator (oracle) as anti-gradients. However, this is a computationally expensive approach that requires many interactions with a physical simulator. One way to accelerate this procedure is to replace the physical simulator with a neural network. Despite recent progress in neural networks for molecular conformation energy prediction, such models are prone to distribution shift, leading to inaccurate energy minimization. We find that the quality of energy minimization with neural networks can be improved by providing optimization trajectories as additional training data. Still, it takes around $5 \times 10^5$ additional conformations to match the physical simulator's optimization quality. In this work, we present the Gradual Optimization Learning Framework (GOLF) for energy minimization with neural networks that significantly reduces the required additional data. The framework consists of an efficient data-collecting scheme and an external optimizer. The external optimizer utilizes gradients from the energy prediction model to generate optimization trajectories, and the data-collecting scheme selects additional training data to be processed by the physical simulator. Our results demonstrate that the neural network trained with GOLF performs on par with the oracle on a benchmark of diverse drug-like molecules using $50$x less additional data.

For agents

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

get_harvested_code_for_paper("2311.06295")
get_code_for_paper("2311.06295")
have("2311.06295")

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

curl https://syntology.ai/api/ran/2311.06295.json

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

[![Syntology run record](https://syntology.ai/api/ran/2311.06295.svg)](https://syntology.ai/paper/2311.06295)

Connect an agent — have() is free.