SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.14997 · ICLR · 2025

Causal Discovery via Bayesian Optimization

Sunil Gupta, Bao Duong, Thin Nguyen

arXiv · PDF · Open in the Atlas

Code that ran

13 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 4 failures of our instrument; 3 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/2501.14997.json · badge: /api/ran/2501.14997.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
baosws/drbo canonical (classifier label)
classifier basis: classifier_2025
9 ran · 3 no recorded run
xunzheng/notears unrelated (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
BaseDagOptim Ran
ran on a synthesized input · constructed an object; no result was computed
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("2069b3643b7b04b6")
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 an instance of its own class, BaseDagOptim: the run constructed the object and did not call it, so no result was computed. 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
from unittest.mock import Mock, MagicMock

# Create a minimal dag_space object with required attributes
class DagSpace:
    def __init__(self):
        self.nodes = 4  # number of nodes in DAG

dag_space = DagSpace()

# Create small numpy arrays for data
X = np.random.randn(10, 4)  # 10 samples, 4 features
GT = np.random.rand(4, 4)   # 4x4 ground truth adjacency matrix

# Create a mock logger with required methods and attributes
logger = Mock()
logger.verbose = False
logger.add = Mock()
logger.__call__ = Mock()

# Create a pruner function that takes X and dag and returns pruned dag
def pruner_func(X, dag):
    return dag

pruner = pruner_func

# Create a scorer function
def scorer_func(X, dag):
    return np.random.rand()

scorer = scorer_func

ARGS = {
    'dag_space': dag_space,
    'X': X,
    'GT': GT,
    'logger': logger,
    'pruner': pruner,
    'scorer': scorer
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
castle, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type BaseDagOptim — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a BaseDagOptim 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
DAGLowRank Ran
ran on a synthesized input · constructed an object; no result was computed
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("3eac5e3dba12300f")
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 an instance of its own class, DAGLowRank: the run constructed the object and did not call it, so no result was computed. 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
from functools import partial
import jax

class DAGLowRank:

    def __init__(self, nodes, rank, **unused):
        self.nodes = nodes
        self.rank = rank
        self.dim = self.nodes * (1 + rank)

    @partial(jax.jit, static_argnums=(0,))
    def __vec2dag(self, z):
        from opt_einsum import contract
        c = z[..., :self.nodes]
        s = z[..., self.nodes:].reshape(*z.shape[-2::-1], self.nodes, self.rank)
        D = contract('...ik,...jk->...ij', s, s)
        C = 1 * (c[..., None] > c[..., None, :])
        E = D > 0
        A = C * E
        return A

    def vec2dag(self, z):
        with jax.default_device(jax.devices('cpu')[0]):
            A = self.__vec2dag(z)
        return np.asarray(A)


nodes = 4
rank = 3

ARGS = {
    "nodes": nodes,
    "rank": rank,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy, opt_einsum — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DAGLowRank — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DAGLowRank 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
LBFGSBScipy Ran
ran on a synthesized input · constructed an object; no result was computed
xunzheng/notears/notears/nonlinear.py
code served (permissive licence) · get_code("8e2af5b396540fc3")
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 an instance of its own class, LBFGSBScipy: the run constructed the object and did not call it, so no result was computed. 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 simple model with parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super(SimpleModel, self).__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = torch.relu(self.linear1(x))
        x = self.linear2(x)
        return x

model = SimpleModel()

# Create some sample data
x_sample = torch.randn(2, 4)
y_sample = torch.randn(2, 2)

# Define closure that computes loss
def create_closure(model, x, y):
    def closure():
        model.zero_grad()
        output = model(x)
        loss = nn.MSELoss()(output, y)
        loss.backward()
        return loss
    return closure

closure = create_closure(model, x_sample, y_sample)

# Prepare params for the optimizer
params = list(model.parameters())

ARGS = {
    'params': params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type LBFGSBScipy — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a LBFGSBScipy 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
Logger Ran
ran on a synthesized input · constructed an object; no result was computed
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("6c20b8e69aa7ebba")
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 an instance of its own class, Logger: the run constructed the object and did not call it, so no result was computed. 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 time

ARGS = {
    'verbose': True
}
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 Logger — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Logger 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
MultiSingleNodeMLP Ran
ran on a synthesized input · constructed an object; no result was computed
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("e4525bd45ca23482")
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 an instance of its own class, MultiSingleNodeMLP: the run constructed the object and did not call it, so no result was computed. 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 torch import nn

torch.manual_seed(0)

ARGS = {
    'nodes': 4,
    'hidden_sizes': [8, 6],
    'batch_norm': True,
    'dropout': 0.1,
    'activation': 'relu',
    'dtype': torch.float32,
}
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 MultiSingleNodeMLP — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MultiSingleNodeMLP 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
Scorer Ran
ran on a synthesized input · constructed an object; no result was computed
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("ce7f72fa60725194")
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 an instance of its own class, Scorer: the run constructed the object and did not call it, so no result was computed. 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

def mock_score_method(X, **params):
    """
    Returns a tuple of (local_score, aggregate, aggregate_batch)
    that the Scorer expects.
    """
    samples, nodes = X.shape
    
    def local_score(node, parents):
        # Simple mock: return a scalar score based on node and parents
        return float(len(parents)) + node * 0.1
    
    def aggregate(dag, scores):
        # Sum all local scores to get overall DAG score
        return float(np.sum(list(scores.values())))
    
    def aggregate_batch(dags, scores_list):
        # Mock batch aggregation
        return np.array([float(np.sum(list(s.values()))) for s in scores_list])
    
    return local_score, aggregate, aggregate_batch

# Create small test data: 5 samples, 4 nodes
data = np.random.rand(5, 4).astype(np.float32)

score_params = {}

ARGS = {
    "score_method": mock_score_method,
    "data": data,
    "score_params": score_params,
}
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 Scorer — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Scorer 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
SingleNodeMLP Ran
ran; our contract draft was wrong, not the code
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("8631372412b39f38")
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 Sequential. 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)

ARGS = {
    'nodes': 4,
    'hidden_sizes': [8, 16],
    'batch_norm': True,
    'dropout': 0.5,
    'activation': 'relu',
    'dtype': torch.float32
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
dual_ascent_step Ran
ran; our contract draft was wrong, not the code
xunzheng/notears/notears/nonlinear.py
code served (permissive licence) · get_code("9234dffd2462bf61")
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 torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

class SimpleModel(nn.Module):
    """Minimal model satisfying the interface needed by dual_ascent_step."""
    
    def __init__(self, input_dim=4):
        super().__init__()
        self.fc1 = nn.Linear(input_dim, 3)
        self.fc2 = nn.Linear(3, input_dim)
    
    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = self.fc2(x)
        return x
    
    def h_func(self):
        """Constraint violation measure - returns a scalar."""
        # Simple constraint: sum of squared weights should not exceed threshold
        w = self.fc1.weight
        return torch.sum(w ** 2) - 0.5
    
    def l2_reg(self):
        """L2 regularization - sum of squared parameters."""
        reg = 0.0
        for param in self.parameters():
            reg = reg + torch.sum(param ** 2)
        return reg
    
    def fc1_l1_reg(self):
        """L1 regularization on fc1 weights."""
        return torch.sum(torch.abs(self.fc1.weight))

# Create model
model = SimpleModel(input_dim=4)

# Create dummy data: batch_size=2, feature_dim=4
X = np.random.randn(2, 4).astype(np.float32)

# Hyperparameters
lambda1 = 0.01
lambda2 = 0.01
rho = 1.0
alpha = 0.0
h = 1.0
rho_max = 100.0

ARGS = {
    'model': model,
    'X': X,
    'lambda1': lambda1,
    'lambda2': lambda2,
    'rho': rho,
    'alpha': alpha,
    'h': h,
    'rho_max': rho_max,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — 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
from_unit_cube Ran
ran, honoured the contract we drafted
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("7f54d13b09d03e90")
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 (2, 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
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# from_unit_cube(x, lb, ub)
# - x: appears to be a tensor that gets multiplied and added
# - lb: lower bound, gets subtracted from ub and multiplied with x
# - ub: upper bound, gets subtracted from lb
# All three should be broadcastable tensors or scalars

# Create small distinct-sized tensors for testing
x = torch.randn(2, 4)  # batch=2, dim=4
lb = torch.randn(2, 4)  # same shape as x
ub = torch.randn(2, 4)  # same shape as x

ARGS = {
    "x": x,
    "lb": lb,
    "ub": ub,
}
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 (2, 4) — shape (2, 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
latin_hypercube Ran
ran, honoured the contract we drafted
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("c846fc2e563e1c26")
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 (5, 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

# Set random seed for reproducibility
np.random.seed(0)

# Define arguments for latin_hypercube function
ARGS = {
    "n_pts": 5,
    "dim": 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 ndarray, ndim 2, dtype family float, shape (5, 3) — shape (5, 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
notears_nonlinear Ran
ran; our contract draft was wrong, not the code
xunzheng/notears/notears/nonlinear.py
code served (permissive licence) · get_code("46e7be5ece660a7c")
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 torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)
np.random.seed(0)

class SimpleDAGModel(nn.Module):
    """Minimal model satisfying the interface used by notears_nonlinear"""
    
    def __init__(self, d: int = 4):
        super().__init__()
        self.d = d
        # fc1 layer for adjacency matrix
        self.fc1 = nn.Linear(d, d, bias=False)
        
    def forward(self, X):
        """Reconstruct X through the DAG model"""
        return self.fc1(X)
    
    def h_func(self):
        """Compute h(W) - the DAG constraint violation"""
        W = self.fc1.weight
        # Frobenius norm based h function (simplified)
        h = torch.trace(torch.matrix_exp(W * W)) - self.d
        return h
    
    def fc1_to_adj(self):
        """Return adjacency matrix as numpy"""
        return self.fc1.weight.data.cpu().detach().numpy()
    
    def l2_reg(self):
        """L2 regularization on fc1 weights"""
        return torch.sum(self.fc1.weight ** 2)
    
    def fc1_l1_reg(self):
        """L1 regularization on fc1 weights"""
        return torch.sum(torch.abs(self.fc1.weight))


# Create model
model = SimpleDAGModel(d=4)

# Create synthetic data: (n_samples=2, n_features=4)
X = np.random.randn(2, 4).astype(np.float32)

ARGS = {
    'model': model,
    'X': X,
    'lambda1': 0.01,
    'lambda2': 0.01,
    'max_iter': 2,
    'h_tol': 1e-8,
    'rho_max': 1e16,
    'w_threshold': 0.3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — 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
squared_loss Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from xunzheng/notears
pointer only · get_code("c60cc034e8957e26")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

to_unit_cube Ran
ran, honoured the contract we drafted
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("8e72d61d7ebc713f")
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 (2, 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
import numpy as np

torch.manual_seed(0)

# Construct arguments for to_unit_cube(x, lb, ub)
# The function normalizes x to [0, 1] range using lower bound (lb) and upper bound (ub)

x = torch.randn(2, 4)  # batch of 2, dimension 4
lb = torch.zeros(4)     # lower bounds for each dimension
ub = torch.ones(4) * 2  # upper bounds for each dimension

ARGS = {
    "x": x,
    "lb": lb,
    "ub": ub
}
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 (2, 4) — shape (2, 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
DagOptimBO Not yet run
unverified (harvested, no recorded run)
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("015f680429c03717")
How it ran

no invocation recorded.

DrBO Not yet run
unverified (harvested, no recorded run)
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("873096d79f18a510")
How it ran

no invocation recorded.

Dropout_Local_BIC Not yet run
unverified (harvested, no recorded run)
baosws/drbo/drbo/drbo.py
code served (permissive licence) · get_code("0c6f285faaba8b51")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 3 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 3 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.

FunctionAlso present inEarliest-datedOthers named
squared_loss
brick of this paper's assembly: loss
8 other papers
7 dated earlier, 1 later · 1 only as an assembly brick
1909.13189 (2019-09)
Learning Sparse Nonparametric DAGs
2105.02522 (2021-05)
Neural graphical modelling in continuous-time: consistency guarantees and algorithms
2106.01474 (2021-06)
Testing Directed Acyclic Graph via Structural, Supervised and Generative Adversarial Learning
and 5 more not named here
LBFGSBScipy
brick of this paper's assembly: scheduler
2 other papers
1 dated earlier, 1 later · 1 only as an assembly brick
2206.06354 (2022-06)
Differentiable and Transportable Structure Learning
2602.23602 (2026-02)
Moment Matters: Mean and Variance Causal Graph Discovery from Heteroscedastic Observational Data
latin_hypercube
brick of this paper's assembly: sampler
1 other paper
1 later
2605.23473 (2026-05)
Automated Random Embedding for Practical Bayesian Optimization with Unknown Effective Dimension

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 score-based methods for directed acyclic graph (DAG) learning from observational data struggle to recover the causal graph accurately and sampleefficiently. To overcome this, in this study, we propose DrBO (DAG recovery via Bayesian Optimization)-a novel DAG learning framework leveraging Bayesian optimization (BO) to find high-scoring DAGs. We show that, by sophisticatedly choosing the promising DAGs to explore, we can find higher-scoring ones much more efficiently. To address the scalability issues of conventional BO in DAG learning, we replace Gaussian Processes commonly employed in BO with dropout neural networks, trained in a continual manner, which allows for (i) flexibly modeling the DAG scores without overfitting, (ii) incorporation of uncertainty into the estimated scores, and (iii) scaling with the number of evaluations. As a result, DrBO is computationally efficient and can find the accurate DAG in fewer trials and less time than existing state-of-the-art methods. This is demonstrated through an extensive set of empirical evaluations on many challenging settings with both synthetic and real data. Our implementation is available at https://github.com/baosws/DrBO. 1 "Sample" in "sample efficiency" refers to the number of trials.

For agents

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

get_harvested_code_for_paper("2501.14997")
get_code_for_paper("2501.14997")
have("2501.14997")

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

curl https://syntology.ai/api/ran/2501.14997.json

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

[![Syntology run record](https://syntology.ai/api/ran/2501.14997.svg)](https://syntology.ai/paper/2501.14997)

Connect an agent — have() is free.