SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.11899 · NeurIPS · 2021

Local policy search with Bayesian optimization

Sarah Müller, Alexander Von Rohr, Sebastian Trimpe

arXiv · PDF · Open in the Atlas

Code that ran

4 of 8 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 4 have no contract verdict; 0 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/2106.11899.json · badge: /api/ran/2106.11899.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
sarmueller/gibo canonical (classifier label)
classifier basis: census
4 ran · 4 no recorded run
FunctionStatusWhere it lives
call_counter Ran
ran on a synthesized input
sarmueller/gibo/src/loop.py
code served (permissive licence) · get_code("5a80e2f4f753b250")
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 function. 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)

def simple_add(x, y):
    """A simple function to be decorated."""
    return x + y

ARGS = {
    "func": simple_add
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
manipulate_reward Ran
ran on a synthesized input
sarmueller/gibo/src/environment_api.py
code served (permissive licence) · get_code("4413d8e3286e6243")
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 function. 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

torch.manual_seed(0)

ARGS = {
    "shift": 0.5,
    "scale": 2.0,
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
optimize_acqf_vanilla_bo Ran
ran on a synthesized input
sarmueller/gibo/src/acquisition_function.py
code served (permissive licence) · get_code("bf9e0cbb23842154")
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 (1, 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 botorch
from botorch.models import SingleTaskGP
from botorch.acquisition import ExpectedImprovement

torch.manual_seed(0)

# Create minimal training data for the GP model
train_x = torch.rand(3, 4)  # 3 observations, 4 dimensions
train_y = torch.rand(3, 1)  # 3 observations, 1 output

# Build a simple GP model
model = SingleTaskGP(train_x, train_y)

# Create an acquisition function (ExpectedImprovement)
best_f = train_y.max().item()
acq_func = ExpectedImprovement(model=model, best_f=best_f)

# Define bounds: 2 x 4 tensor (lower and upper bounds for 4 dimensions)
bounds = torch.tensor([[0.0, 0.0, 0.0, 0.0], 
                       [1.0, 1.0, 1.0, 1.0]])

ARGS = {
    "acq_func": acq_func,
    "bounds": bounds,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
botorch, 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 (1, 4) — shape (1, 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
sort_rewards_global_optimization Ran
ran on a synthesized input
sarmueller/gibo/src/evaluate.py
code served (permissive licence) · get_code("9493a6a3cdcb4b44")
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 copy
import numpy as np

# Construct the rewards dictionary
# The function expects: dict mapping dimension -> list of objective rewards
# Each objective reward is a list of reward values
rewards = {
    0: [
        [1.0, 0.5, 2.0, 1.5, 3.0],
        [0.2, 0.8, 0.5, 1.2, 0.9]
    ],
    1: [
        [2.5, 1.0, 3.5, 2.0, 4.0],
        [0.1, 0.7, 0.3, 1.0, 0.6]
    ]
}

ARGS = {
    "rewards": rewards
}
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
insert Not yet run
unverified (harvested, no recorded run)
sarmueller/gibo/src/config.py
code served (permissive licence) · get_code("9ba1bb4104c6cbdc")
How it ran

no invocation recorded.

one_step_cholesky Not yet run
unverified (harvested, no recorded run)
sarmueller/gibo/src/cholesky.py
code served (permissive licence) · get_code("df94ac83c17a7fb3")
How it ran

no invocation recorded.

optimize_acqf_custom_bo Not yet run
unverified (harvested, no recorded run)
sarmueller/gibo/src/acquisition_function.py
code served (permissive licence) · get_code("20307dfa91e06968")
How it ran

no invocation recorded.

postprocessing_interpolate_x Not yet run
unverified (harvested, no recorded run)
sarmueller/gibo/src/evaluate.py
code served (permissive licence) · get_code("93332fe605a1eec4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 8 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 8 later) and 0 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
call_counter 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
insert 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
manipulate_reward 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
one_step_cholesky 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
optimize_acqf_custom_bo 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
optimize_acqf_vanilla_bo 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
postprocessing_interpolate_x 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent
sort_rewards_global_optimization 1 other paper
1 later
2210.11662 (2022-10)
Local Bayesian optimization via maximizing probability of descent

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

Reinforcement learning (RL) aims to find an optimal policy by interaction with an environment. Consequently, learning complex behavior requires a vast number of samples, which can be prohibitive in practice. Nevertheless, instead of systematically reasoning and actively choosing informative samples, policy gradients for local search are often obtained from random perturbations. These random samples yield high variance estimates and hence are sub-optimal in terms of sample complexity. Actively selecting informative samples is at the core of Bayesian optimization, which constructs a probabilistic surrogate of the objective from past samples to reason about informative subsequent ones. In this paper, we propose to join both worlds. We develop an algorithm utilizing a probabilistic model of the objective function and its gradient. Based on the model, the algorithm decides where to query a noisy zeroth-order oracle to improve the gradient estimates. The resulting algorithm is a novel type of policy search method, which we compare to existing black-box algorithms. The comparison reveals improved sample complexity and reduced variance in extensive empirical evaluations on synthetic objectives. Further, we highlight the benefits of active sampling on popular RL benchmarks.

For agents

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

get_harvested_code_for_paper("2106.11899")
get_code_for_paper("2106.11899")
have("2106.11899")

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

curl https://syntology.ai/api/ran/2106.11899.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.11899.svg)](https://syntology.ai/paper/2106.11899)

Connect an agent — have() is free.