SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.03113 · 2020

Projective Preferential Bayesian Optimization

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-21; 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/2002.03113.json · badge: /api/ran/2002.03113.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
P-Mikkola/PPBO community (classifier label)
classifier basis not recorded on the link
7 ran · 2 no recorded run
FunctionStatusWhere it lives
EI Ran
ran on a synthesized input
P-Mikkola/PPBO/src/acquisition.py
code served (permissive licence) · get_code("bbf4831cff62777c")
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 float64, ndim 0, 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 numpy as np
import torch
import torch.nn as nn

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

class StandInGPModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.mustar = 0.5
        self.FP = StandInFP()
    
    def mu_Sigma_pred(self, xi_grid):
        # xi_grid shape: (m, d)
        # Return mean vector and covariance matrix
        m = xi_grid.shape[0]
        f_post_mean = np.random.randn(m)
        f_post_covar = np.eye(m) * 0.1 + np.ones((m, m)) * 0.01
        return f_post_mean, f_post_covar

class StandInFP:
    def xi_grid(self, xi, x, alpha_grid_distribution, alpha_star, m, is_scaled):
        # Return a grid of shape (m, d)
        # d=4 for consistency with x
        return np.random.randn(m, 4)

# Create instances
gp_model = StandInGPModel()
xi = np.random.randn(4)
x = np.random.randn(3, 4)
mc_samples = 5

ARGS = {
    'xi': xi,
    'x': x,
    'GP_model': gp_model,
    'mc_samples': mc_samples
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
EI_to_maximize Ran
ran on a synthesized input
P-Mikkola/PPBO/src/acquisition.py
code served (permissive licence) · get_code("03f203f5d55f326b")
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 float64, ndim 0, 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 numpy as np
import torch
import torch.nn as nn

class StandInGPModel(nn.Module):
    def __init__(self, D=5, xi_grid_size=70):
        super().__init__()
        self.D = D
        self.xi_grid_size = xi_grid_size
        self.mustar = 2.5
        self.FP = StandInFP()
        
    def xi_grid(self, xi, x, alpha_grid_distribution, alpha_star, m, is_scaled):
        # Return a grid of points
        return np.random.randn(m, self.D)
    
    def mu_Sigma_pred(self, xi_grid):
        # Return mean and covariance for the grid
        m = xi_grid.shape[0]
        f_post_mean = np.random.randn(m)
        # Create a valid positive-definite covariance matrix
        A = np.random.randn(m, m)
        f_post_covar = A @ A.T + np.eye(m) * 0.1
        return f_post_mean, f_post_covar

class StandInFP:
    def xi_grid(self, xi, x, alpha_grid_distribution, alpha_star, m, is_scaled):
        # Return a grid of points
        D = len(xi)
        return np.random.randn(m, D)

# Create the GP model
gp_model = StandInGPModel(D=5)

# Set up arguments
xi_plus_x_value = np.array([0.1, 0.2, 0.3, 0.4, 0.5])
xi_dims = np.array([0, 1, 2])
x_dims = np.array([3, 4])
mc_samples_value = 3

ARGS = {
    'xi_plus_x': np.array([xi_plus_x_value]),
    'xi_dims': xi_dims,
    'x_dims': x_dims,
    'GP_model': gp_model,
    'mc_samples': mc_samples_value
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
SE_kernel Ran
ran on a synthesized input
P-Mikkola/PPBO/src/kernels.py
code served (permissive licence) · get_code("d1bf8cda6854aa33")
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 (2, 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

# Create sample data for X1 and X2
X1 = np.array([[1.0, 2.0, 3.0, 4.0],
               [5.0, 6.0, 7.0, 8.0]])

X2 = np.array([[0.5, 1.5, 2.5, 3.5],
               [4.5, 5.5, 6.5, 7.5],
               [9.0, 10.0, 11.0, 12.0]])

# Create hyperparameter theta
# theta[0] is typically sigma_n (noise), theta[1] is l (length scale), theta[2] is sigma_f (signal variance)
theta = np.array([0.1, 1.0, 2.0])

ARGS = {
    'X1': X1,
    'X2': X2,
    'theta': theta
}
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 (2, 3) — shape (2, 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
alpha_bounds Ran
ran on a synthesized input
P-Mikkola/PPBO/src/misc.py
code served (permissive licence) · get_code("d2966880e7b5a6db")
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 numpy as np

# Construct valid arguments for alpha_bounds function
# The function expects:
# - xi: array-like, direction vector (can have positive, negative, and zero values)
# - lower: array-like, lower bounds
# - upper: array-like, upper bounds

# Create a scenario where we have some meaningful constraints
# xi represents a direction vector with mixed signs
xi = np.array([2.0, -1.5, 3.0, -0.5])

# lower and upper bounds for each dimension
lower = np.array([1.0, -3.0, 2.0, -2.0])
upper = np.array([5.0, -1.0, 8.0, 0.0])

ARGS = {
    "xi": xi,
    "lower": lower,
    "upper": upper
}
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
d Ran
ran on a synthesized input
P-Mikkola/PPBO/src/kernels.py
code served (permissive licence) · get_code("e2ec08ac02d66882")
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, 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

# The function d(x1, x2) computes absolute distance between all elements
# of two vectors using np.subtract.outer(), which broadcasts the subtraction
# and returns a 2D matrix of absolute differences.
# 
# Parameters:
# - x1: a 1D array-like (vector)
# - x2: a 1D array-like (vector)

# Create two small 1D numpy arrays with distinct sizes
x1 = np.array([1.0, 2.0, 3.0, 4.0])  # size 4
x2 = np.array([5.0, 6.0, 7.0])       # size 3

ARGS = {
    "x1": x1,
    "x2": x2,
}
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 (4, 3) — shape (4, 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
dist Ran
ran on a synthesized input
P-Mikkola/PPBO/src/kernels.py
code served (permissive licence) · get_code("569567299c57a97c")
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 (2, 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

# Create two small numpy arrays with distinct dimensions
# X1: 2 rows (samples), 4 columns (features)
# X2: 3 rows (samples), 4 columns (features)
X1 = np.array([[1.0, 2.0, 3.0, 4.0],
               [5.0, 6.0, 7.0, 8.0]], dtype=np.float32)

X2 = np.array([[1.5, 2.5, 3.5, 4.5],
               [5.5, 6.5, 7.5, 8.5],
               [9.0, 10.0, 11.0, 12.0]], dtype=np.float32)

ARGS = {
    'X1': X1,
    'X2': X2
}
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 (2, 3) — shape (2, 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
regularize_covariance Ran
ran on a synthesized input
P-Mikkola/PPBO/src/misc.py
code served (permissive licence) · get_code("9ba9ae60d283b6b8")
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 numpy as np
import sklearn.covariance

# Create a small covariance matrix for testing
# Using a 4x4 matrix to keep it small but realistic
np.random.seed(42)
X = np.random.randn(10, 4)
cov_matrix = np.cov(X.T)  # 4x4 covariance matrix

ARGS = {
    "X": cov_matrix,
    "reg_level": 0.0001,
    "pos_diag": True,
    "jitter": 1e-07,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
inverse Not yet run
unverified (harvested, no recorded run)
P-Mikkola/PPBO/src/misc.py
code served (permissive licence) · get_code("31352f1d9134ca1b")
How it ran

no invocation recorded.

sliceplot_pred_mean Not yet run
unverified (harvested, no recorded run)
P-Mikkola/PPBO/camphor_copper/plot_results.py
code served (permissive licence) · get_code("856f4852848eb8c5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 9 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

Repositories linked to this paper

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

Abstract

Bayesian optimization is an effective method for finding extrema of a black-box function. We propose a new type of Bayesian optimization for learning user preferences in high-dimensional spaces. The central assumption is that the underlying objective function cannot be evaluated directly, but instead a minimizer along a projection can be queried, which we call a projective preferential query. The form of the query allows for feedback that is natural for a human to give, and which enables interaction. This is demonstrated in a user experiment in which the user feedback comes in the form of optimal position and orientation of a molecule adsorbing to a surface. We demonstrate that our framework is able to find a global minimum of a high-dimensional black-box function, which is an infeasible task for existing preferential Bayesian optimization frameworks that are based on pairwise comparisons.

For agents

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

get_harvested_code_for_paper("2002.03113")
get_code_for_paper("2002.03113")
have("2002.03113")

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

curl https://syntology.ai/api/ran/2002.03113.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.03113.svg)](https://syntology.ai/paper/2002.03113)

Connect an agent — have() is free.