SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.20258 · 2023

Advancing Bayesian Optimization via Learning Correlated Latent Space

arXiv · PDF · Open in the Atlas

Code that ran

6 of 10 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 4 failures of our instrument; 4 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/2310.20258.json · badge: /api/ran/2310.20258.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
mlvlab/CoBO canonical (classifier label)
classifier basis: census
6 ran · 4 no recorded run
FunctionStatusWhere it lives
compute_train_zs Ran
ran; our contract draft was wrong, not the code
mlvlab/CoBO/cobo/utils/mol_utils/load_data.py
code served (permissive licence) · get_code("61a73be28a3229ec")
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 (10, 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 math

torch.manual_seed(0)

class SimpleVAE(nn.Module):
    def __init__(self, latent_dim=4, input_dim=8):
        super().__init__()
        self.latent_dim = latent_dim
        self.input_dim = input_dim
        self.encoder = nn.Linear(input_dim, latent_dim * 2)
        self.decoder = nn.Linear(latent_dim, input_dim)
    
    def forward(self, x):
        h = self.encoder(x)
        mu, logvar = h.chunk(2, dim=1)
        std = torch.exp(0.5 * logvar)
        eps = torch.randn_like(std)
        z = mu + eps * std
        return z, mu, logvar, None, None

class MolObjective:
    def __init__(self, vae, path_to_vae_statedict):
        self.vae = vae
        self.path_to_vae_statedict = path_to_vae_statedict
    
    def vae_forward(self, xs_batch):
        return self.vae(xs_batch)

vae = SimpleVAE(latent_dim=4, input_dim=8)
vae.eval()

mol_objective = MolObjective(
    vae=vae,
    path_to_vae_statedict="/tmp/vae_model.pt"
)

init_train_x = torch.randn(10, 8)

ARGS = {
    "mol_objective": mol_objective,
    "init_train_x": init_train_x,
    "bsz": 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, 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 (10, 4) — shape (10, 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
gumbel_softmax Ran
ran; our contract draft was wrong, not the code
mlvlab/CoBO/cobo/utils/mol_utils/selfies_vae/model_positional_unbounded.py
code served (permissive licence) · get_code("5b05e8e843deabaf")
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
from torch import Tensor

torch.manual_seed(0)

# Create a small logits tensor (batch=2, num_classes=4)
logits = torch.randn(2, 4)

# Create optional randoms tensor with same shape as logits
randoms = -torch.empty_like(logits, memory_format=torch.legacy_contiguous_format).exponential_().log()

ARGS = {
    "logits": logits,
    "tau": 1.0,
    "hard": False,
    "dim": -1,
    "return_randoms": False,
    "randoms": randoms,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 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
is_valid_molecule Ran
ran, violated the contract we drafted
mlvlab/CoBO/cobo/utils/mol_utils/selfies_vae/model_positional_unbounded.py
code served (permissive licence) · get_code("f8e7af52c159f99d")
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 torch.nn as nn

torch.manual_seed(0)

# The function is_valid_molecule takes a single parameter x
# and returns True. Since the function body is trivial (just returns True),
# x can be any value. We'll use a simple molecule representation.

# A molecule could be represented as:
# - A string (SMILES notation)
# - A tensor (feature representation)
# - A dictionary (molecular properties)
# - A custom object

# Since the function doesn't actually use x in any way (just returns True),
# we can pass any valid Python object. Let's use a few reasonable options:

# Option 1: SMILES string representation
molecule_smiles = "CCO"  # ethanol

# Option 2: Tensor representation (molecular features)
molecule_tensor = torch.randn(2, 4)

# Option 3: Dictionary representation
molecule_dict = {
    "atoms": ["C", "C", "O"],
    "bonds": [(0, 1), (1, 2)],
    "features": torch.randn(2, 4)
}

# Option 4: Simple custom class
class Molecule:
    def __init__(self):
        self.atoms = ["C", "H", "O"]
        self.features = torch.randn(2, 4)

molecule_obj = Molecule()

# Since the function just returns True regardless of input,
# we can use any of these. Let's use a simple SMILES string.
ARGS = {
    "x": "CCO"
}
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
lipschitz_loss Ran
ran; our contract draft was wrong, not the code
mlvlab/CoBO/cobo/utils/utils.py
code served (permissive licence) · get_code("98cfe3d3a797ae62")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# z: tensor that gets .cuda() called, then unsqueezed and used in distance computation
# Shape: (batch_size, embedding_dim)
z = torch.randn(2, 4)

# y: tensor that gets .cuda() called, then unsqueezed and squeezed
# Shape: (batch_size, 1) based on squeeze(-1) operation
y = torch.randn(2, 1)

# recon_weight: tensor used in multiplication and indexing with [:, None]
# Shape: (batch_size,)
recon_weight = torch.randn(2)

ARGS = {
    "z": z,
    "y": y,
    "recon_weight": recon_weight,
}
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, 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
load_train_z Ran
ran, honoured the contract we drafted
mlvlab/CoBO/cobo/utils/mol_utils/load_data.py
code served (permissive licence) · get_code("e00ccb61913dc9fb")
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 (5, 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 pandas as pd
import torch
import tempfile
import os

torch.manual_seed(0)

# Create a temporary directory and CSV file for testing
temp_dir = tempfile.mkdtemp()
csv_path = os.path.join(temp_dir, "model-train-zs.csv")

# Create sample training z values
sample_zs = torch.randn(10, 4)
df = pd.DataFrame(sample_zs.numpy())
df.to_csv(csv_path, header=False, index=False)

# Path to VAE state dict (will be inferred from csv path)
path_to_vae_statedict = csv_path.replace("-train-zs.csv", ".pt")

ARGS = {
    "num_initialization_points": 5,
    "path_to_vae_statedict": path_to_vae_statedict,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, 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 (5, 4) — shape (5, 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
update_state Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mlvlab/CoBO/cobo/utils/bo_utils/turbo.py
code served (permissive licence) · get_code("ba51f9dac22959af")
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 State. 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 math

torch.manual_seed(0)

class State:
    def __init__(self):
        self.best_value = 0.5
        self.success_counter = 0
        self.failure_counter = 0
        self.success_tolerance = 3
        self.failure_tolerance = 2
        self.length = 1.0
        self.length_max = 10.0
        self.length_min = 0.01
        self.restart_triggered = False

state = State()
Y_next = torch.tensor([0.7, 0.9])

ARGS = {
    'state': state,
    'Y_next': Y_next
}
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 State — a State output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a State 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
generate_batch Not yet run
unverified (harvested, no recorded run)
mlvlab/CoBO/cobo/utils/bo_utils/turbo.py
code served (permissive licence) · get_code("a1d32b710e32cd21")
How it ran

no invocation recorded.

load_molecule_train_data Not yet run
unverified (harvested, no recorded run)
mlvlab/CoBO/cobo/utils/mol_utils/load_data.py
code served (permissive licence) · get_code("dc3774d0bdbe2e6b")
How it ran

no invocation recorded.

update_models_end_to_end Not yet run
unverified (harvested, no recorded run)
mlvlab/CoBO/cobo/utils/utils.py
code served (permissive licence) · get_code("75d4f55ca9e56683")
How it ran

no invocation recorded.

update_surr_model Not yet run
unverified (harvested, no recorded run)
mlvlab/CoBO/cobo/utils/utils.py
code served (permissive licence) · get_code("0dda692b10a19709")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 6 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 4 later) and 4 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
update_state 3 other papers
1 dated earlier, 2 later
2201.11872 (2022-01)
Local Latent Space Bayesian Optimization over Structured Inputs
2406.05250 (2024-06)
LLM-Enhanced Bayesian Optimization for Efficient Analog Layout Constraint Generation
2411.05330 (2024-11)
Inversion-based Latent Bayesian Optimization
load_train_z 2 other papers
1 dated earlier, 1 later
2201.11872 (2022-01)
Local Latent Space Bayesian Optimization over Structured Inputs
2411.05330 (2024-11)
Inversion-based Latent Bayesian Optimization
update_surr_model 2 other papers
1 dated earlier, 1 later
2201.11872 (2022-01)
Local Latent Space Bayesian Optimization over Structured Inputs
2411.05330 (2024-11)
Inversion-based Latent Bayesian Optimization
gumbel_softmax 1 other paper
1 dated earlier
2201.11872 (2022-01)
Local Latent Space Bayesian Optimization over Structured Inputs
is_valid_molecule 1 other paper
1 dated earlier
2201.11872 (2022-01)
Local Latent Space Bayesian Optimization over Structured Inputs
lipschitz_loss 1 other paper
1 later
2411.05330 (2024-11)
Inversion-based Latent Bayesian Optimization

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

Bayesian optimization is a powerful method for optimizing black-box functions with limited function evaluations. Recent works have shown that optimization in a latent space through deep generative models such as variational autoencoders leads to effective and efficient Bayesian optimization for structured or discrete data. However, as the optimization does not take place in the input space, it leads to an inherent gap that results in potentially suboptimal solutions. To alleviate the discrepancy, we propose Correlated latent space Bayesian Optimization (CoBO), which focuses on learning correlated latent spaces characterized by a strong correlation between the distances in the latent space and the distances within the objective function. Specifically, our method introduces Lipschitz regularization, loss weighting, and trust region recoordination to minimize the inherent gap around the promising areas. We demonstrate the effectiveness of our approach on several optimization tasks in discrete data, such as molecule design and arithmetic expression fitting, and achieve high performance within a small budget.

For agents

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

get_harvested_code_for_paper("2310.20258")
get_code_for_paper("2310.20258")
have("2310.20258")

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

curl https://syntology.ai/api/ran/2310.20258.json

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

[![Syntology run record](https://syntology.ai/api/ran/2310.20258.svg)](https://syntology.ai/paper/2310.20258)

Connect an agent — have() is free.