SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2201.00649 · NeurIPS · 2021

SAE: Sequential Anchored Ensembles

Gilles Louppe, Arnaud Delaunoy

arXiv · PDF · Open in the Atlas

Code that ran

9 of 12 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 3 violated it and 4 have no contract verdict; 2 failures of our instrument; 3 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/2201.00649.json · badge: /api/ran/2201.00649.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
adelau/sae-sequential-anchored-ensembles canonical (classifier label)
classifier basis: classifier_2025
9 ran · 3 no recorded run
FunctionStatusWhere it lives
GaussianMHSampler Ran
ran on a synthesized input · constructed an object; no result was computed
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("61f6c4b092d7b700")
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, GaussianMHSampler: 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 scipy.stats import norm

# Set random seed for reproducibility
np.random.seed(42)

# Define arguments for GaussianMHSampler
means = np.array([0.0, 1.0, 2.0, 3.0])
stds = np.array([0.5, 1.0, 1.5, 2.0])
step_std = 0.3

ARGS = {
    "means": means,
    "stds": stds,
    "step_std": step_std,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GaussianMHSampler — 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 GaussianMHSampler 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
GuidedWalkGaussianSampler Ran
ran on a synthesized input · constructed an object; no result was computed
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("6b8f48f671cd0837")
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, GuidedWalkGaussianSampler: 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 scipy.stats import norm

# Set random seed for reproducibility
np.random.seed(42)

# Define small test arrays for the sampler
means = np.array([0.0, 1.0, -0.5, 2.0])
stds = np.array([1.0, 0.5, 2.0, 1.5])
step_std = 0.1

ARGS = {
    "means": means,
    "stds": stds,
    "step_std": step_std,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GuidedWalkGaussianSampler — 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 GuidedWalkGaussianSampler 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
HamiltonianSampler Ran
ran on a synthesized input · constructed an object; no result was computed
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("9d0d07fa43a7436f")
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, HamiltonianSampler: 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

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

# Create small stand-in arrays for the HamiltonianSampler constructor
means = np.array([0.0, 0.5, -0.3, 0.2])  # 4 dimensions
stds = np.array([1.0, 0.8, 1.2, 0.9])    # 4 dimensions, distinct from means
step_size = 0.1
num_steps = 5

ARGS = {
    "means": means,
    "stds": stds,
    "step_size": step_size,
    "num_steps": num_steps,
}
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 HamiltonianSampler — 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 HamiltonianSampler 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
IIDSampler Ran
ran on a synthesized input · constructed an object; no result was computed
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("bb8acbf1d5ebc0eb")
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, IIDSampler: 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

np.random.seed(0)

# Create simple numpy arrays for means and stds
# Using distinct sizes to make shapes unambiguous
means = np.array([0.0, 1.0, 2.0, 3.0])
stds = np.array([0.1, 0.2, 0.3, 0.4])

ARGS = {
    "means": means,
    "stds": stds,
}
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 IIDSampler — 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 IIDSampler 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
gaussian_energy Ran
ran, violated the contract we drafted
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("89bd65f408cfe178")
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 1, dtype family float, shape (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 math
import numpy as np
import torch

torch.manual_seed(0)

# Analyze function usage:
# - x: used with len(x), arithmetic operations → needs to be array-like with length
# - means: used in arithmetic with x → scalar or array
# - sigmas: used in arithmetic with x → scalar or array

# Create small test arrays
x = np.array([0.5, 1.2, -0.3, 0.8])
means = 0.0
sigmas = 1.0

ARGS = {
    "x": x,
    "means": means,
    "sigmas": sigmas,
}
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 ndarray, ndim 1, dtype family float, shape (4,) — shape (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
gaussian_energy_grad Ran
ran, violated the contract we drafted
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("32cce0754db5f177")
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 1, dtype family float, shape (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

ARGS = {
    "x": np.array([1.0, 2.0, 3.0, -1.0]),
    "means": np.array([0.0, 0.5, 1.0, -0.5]),
    "sigmas": np.array([1.0, 1.5, 2.0, 0.8])
}
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 1, dtype family float, shape (4,) — shape (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
hamiltonian Ran
ran; our contract draft was wrong, not the code
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("b74debc066a4106d")
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, 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

# Define a simple energy function that takes position and returns energy
def energy_function(position):
    """Simple quadratic energy function"""
    return 0.5 * np.sum(position ** 2, axis=-1, keepdims=True)

# Create small test arrays
position = np.random.randn(2, 4).astype(np.float32)
velocity = np.random.randn(2, 4).astype(np.float32)

ARGS = {
    "position": position,
    "velocity": velocity,
    "energy": energy_function
}
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, 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
kinetic_energy Ran
ran, violated the contract we drafted
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("02d422947f4c8763")
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 1, dtype family float, shape (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

# Create a small velocity array
velocity = np.array([1.0, 2.0, 3.0, 4.0])

ARGS = {
    "velocity": velocity
}
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 1, dtype family float, shape (4,) — shape (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
leapfrog_step Ran
ran; our contract draft was wrong, not the code
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("ed00346c5673fc9f")
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 numpy as np
import torch

torch.manual_seed(0)

# Define a simple energy function
def energy_fn(x):
    """Simple quadratic energy function"""
    return 0.5 * np.sum(x ** 2)

# Define the gradient of the energy function
def grad_energy_fn(x):
    """Gradient of the quadratic energy function"""
    return x.copy()

# Create initial positions and velocities
x0 = np.array([1.0, 2.0, 3.0, 0.5])
v0 = np.array([0.1, 0.2, 0.15, 0.05])

# Define step size and number of steps
step_size = 0.01
num_steps = 5

ARGS = {
    "x0": x0,
    "v0": v0,
    "energy": energy_fn,
    "grad_energy": grad_energy_fn,
    "step_size": step_size,
    "num_steps": num_steps,
}
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 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
SequentialEnsembleModel Not yet run
unverified (harvested, no recorded run)
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("af37d6c827e3f20b")
How it ran

no invocation recorded.

hmc Not yet run
unverified (harvested, no recorded run)
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("d086c69c29f38bbe")
How it ran

no invocation recorded.

train_model Not yet run
unverified (harvested, no recorded run)
adelau/sae-sequential-anchored-ensembles/models/models.py
code served (permissive licence) · get_code("b5119a1ea551b8f5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 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 12 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

Computing the Bayesian posterior of a neural network is a challenging task due to the high-dimensionality of the parameter space. Anchored ensembles approximate the posterior by training an ensemble of neural networks on anchored losses designed for the optima to follow the Bayesian posterior. Training an ensemble, however, becomes computationally expensive as its number of members grows since the full training procedure is repeated for each member. In this note, we present Sequential Anchored Ensembles (SAE), a lightweight alternative to anchored ensembles. Instead of training each member of the ensemble from scratch, the members are trained sequentially on losses sampled with high auto-correlation, hence enabling fast convergence of the neural networks and efficient approximation of the Bayesian posterior. SAE outperform anchored ensembles, for a given computational budget, on some benchmarks while showing comparable performance on the others and achieved 2 nd and 3 rd place in the light and extended tracks of the NeurIPS 2021 Approximate Inference in Bayesian Deep Learning competition.

For agents

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

get_harvested_code_for_paper("2201.00649")
get_code_for_paper("2201.00649")
have("2201.00649")

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

curl https://syntology.ai/api/ran/2201.00649.json

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

[![Syntology run record](https://syntology.ai/api/ran/2201.00649.svg)](https://syntology.ai/paper/2201.00649)

Connect an agent — have() is free.