SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1612.01474 · 2016

Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles

Charles Blundell, Alexander Pritzel, Balaji Lakshminarayanan

arXiv · PDF · Open in the Atlas

Code that ran

10 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (8 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 9 have no contract verdict; 1 failure of our instrument; 1 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/1612.01474.json · badge: /api/ran/1612.01474.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
ENSTA-U2IS-AI/torch-uncertainty community
Papers with Code archive
3 ran · 0 no recorded run
Junghwan-brian/SDE-Net community
Papers with Code archive
2 ran · 1 no recorded run
JoonHyung-Park/DeepEnsemble community
Papers with Code archive
2 ran · 0 no recorded run
StanfordASL/SCOD community
Papers with Code archive
1 ran · 0 no recorded run
github-jnauta/pytorch-pne community
Papers with Code archive
1 ran · 0 no recorded run
tkerscher/blast community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
Diffusion Ran
ran on a synthesized input · constructed an object; no result was computed
Junghwan-brian/SDE-Net/model/SDENet.py
pointer only (licence: NONE) · get_code("59b274fed1c2f7d5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Junghwan-brian/SDE-Net when it was checked. 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.

Drift Ran
ran on a synthesized input · constructed an object; no result was computed
Junghwan-brian/SDE-Net/model/SDENet.py
pointer only (licence: NONE) · get_code("9ceff01e171db846")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Junghwan-brian/SDE-Net when it was checked. 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.

Ensemble Ran
ran on a synthesized input · constructed an object; no result was computed
StanfordASL/SCOD/nn_ood/posteriors/ensemble.py
code served (permissive licence) · get_code("cfe90024d387377b")
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, Ensemble: the run constructed the object and did not call it, so no result was computed. 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 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)

# Create a simple model that takes input and outputs 1-dimensional output
class SimpleModel(nn.Module):
    def __init__(self, input_dim=4, output_dim=1):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return self.linear(x)

# Create multiple trained models (just different instances with different weights)
model1 = SimpleModel(input_dim=4, output_dim=1)
model2 = SimpleModel(input_dim=4, output_dim=1)
model3 = SimpleModel(input_dim=4, output_dim=1)

models = [model1, model2, model3]

# Create a distribution family class with merge_ensemble method
class DistributionFamily:
    def merge_ensemble(self, outputs):
        """
        Takes stacked outputs of shape (n_models, batch_size, output_dim)
        Returns mu (mean) and unc (uncertainty)
        """
        # outputs shape: (n_models, batch_size, 1)
        mu = torch.mean(outputs, dim=0)  # shape: (batch_size, 1)
        # uncertainty as std across models
        unc = torch.std(outputs, dim=0).squeeze(-1)  # shape: (batch_size,)
        return mu, unc

dist_fam = DistributionFamily()

# Create args dict
args = {'device': 'cpu'}

ARGS = {
    'models': models,
    'dist_fam': dist_fam,
    'args': args
}
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 Ensemble — 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 Ensemble 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
GaussianMLP Ran
ran on a synthesized input · constructed an object; no result was computed
JoonHyung-Park/DeepEnsemble/complete_ver/model.py
pointer only (licence: NONE) · get_code("2bb0022c0652839d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, JoonHyung-Park/DeepEnsemble when it was checked. 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.

MLP Ran
ran on a synthesized input · constructed an object; no result was computed
JoonHyung-Park/DeepEnsemble/complete_ver/model.py
pointer only (licence: NONE) · get_code("8f0ebe8c1a69cb4b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, JoonHyung-Park/DeepEnsemble when it was checked. 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.

Model Ran
ran on a synthesized input · constructed an object; no result was computed
github-jnauta/pytorch-pne/models/pnn.py
pointer only (licence: NONE) · get_code("1fbc81b2bd6ea48b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, github-jnauta/pytorch-pne when it was checked. 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.

_DeepEnsembles Ran
ran on a synthesized input · constructed an object; no result was computed
ENSTA-U2IS-AI/torch-uncertainty/src/torch_uncertainty/methods/deep_ensembles.py
code served (permissive licence) · get_code("b8e6a0ddc977191d")
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, _DeepEnsembles: 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 small stand-in models for the ensemble
class SimpleModel(nn.Module):
    def __init__(self, input_dim: int = 4, output_dim: int = 3):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x: torch.Tensor) -> torch.Tensor:
        return self.linear(x)

# Create a list of core models
core_models = [
    SimpleModel(input_dim=4, output_dim=3),
    SimpleModel(input_dim=4, output_dim=3),
    SimpleModel(input_dim=4, output_dim=3),
]

ARGS = {
    "core_models": core_models,
    "store_on_cpu": False,
}
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 _DeepEnsembles — 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 _DeepEnsembles 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
_RegDeepEnsembles Ran
ran on a synthesized input · constructed an object; no result was computed
ENSTA-U2IS-AI/torch-uncertainty/src/torch_uncertainty/methods/deep_ensembles.py
code served (permissive licence) · get_code("5425e2595166ea3a")
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, _RegDeepEnsembles: 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)

# Create small dummy models for the ensemble
class DummyRegressionModel(nn.Module):
    def __init__(self, output_keys=None, probabilistic=False):
        super().__init__()
        self.linear = nn.Linear(4, 3)
        self.probabilistic = probabilistic
        self.output_keys = output_keys or ['mean', 'std']
    
    def forward(self, x):
        out = self.linear(x)
        if self.probabilistic:
            # Return dict with keys for probabilistic output
            return {key: out for key in self.output_keys}
        return out

# Create core models for the ensemble
model1 = DummyRegressionModel(probabilistic=True)
model2 = DummyRegressionModel(probabilistic=True)
model3 = DummyRegressionModel(probabilistic=True)

core_models = [model1, model2, model3]

ARGS = {
    'probabilistic': True,
    'core_models': core_models,
    'store_on_cpu': False
}
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 _RegDeepEnsembles — 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 _RegDeepEnsembles 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
deep_ensembles Ran
ran; our contract draft was wrong, not the code
ENSTA-U2IS-AI/torch-uncertainty/src/torch_uncertainty/methods/deep_ensembles.py
code served (permissive licence) · get_code("a76c86e0e721e4b2")
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) and returned _DeepEnsembles. 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
from pathlib import Path

torch.manual_seed(0)

# Create a simple model for testing
class SimpleModel(nn.Module):
    def __init__(self, input_dim=4, output_dim=3):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return self.linear(x)

class SimpleProbabilisticModel(nn.Module):
    def __init__(self, input_dim=4, output_dim=3):
        super().__init__()
        self.linear_mean = nn.Linear(input_dim, output_dim)
        self.linear_std = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return {
            'mean': self.linear_mean(x),
            'std': torch.nn.functional.softplus(self.linear_std(x))
        }

# Create core models for the ensemble
core_model_1 = SimpleModel(input_dim=4, output_dim=3)
core_model_2 = SimpleModel(input_dim=4, output_dim=3)
core_model_3 = SimpleModel(input_dim=4, output_dim=3)

core_models_list = [core_model_1, core_model_2, core_model_3]

ARGS = {
    'core_models': core_models_list,
    'num_estimators': None,
    'task': 'classification',
    'probabilistic': None,
    'reset_model_parameters': True,
    'store_on_cpu': False,
    'ckpt_paths': None,
    'use_tu_ckpt_format': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _DeepEnsembles — a _DeepEnsembles output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _DeepEnsembles 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
parse_sed Ran
ran on a synthesized input
tkerscher/blast/blast/parser.py
code served (permissive licence) · get_code("a03d09d6f461b685")
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 (3, 2) 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
from io import StringIO

# Create a valid SED string with proper format
# Format: lines with at least 7 entries, where entry[6] == 'Det'
# First line should have at least 5 entries (for position parsing)
sed_content = """# Object RA Dec 10.5 20.3
# Some header
# Another header
# Column info
1.0 2.0 2.5 1.5 0.0 0.0 Det
10.0 3.0 3.5 2.5 0.0 0.0 Det
100.0 4.0 4.5 3.5 0.0 0.0 Det
"""

ARGS = {
    "sed": sed_content,
    "sanitize": True,
    "position": False
}
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 (3, 2) — shape (3, 2) 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
SDENet Not yet run
unverified (harvested, no recorded run)
Junghwan-brian/SDE-Net/model/SDENet.py
pointer only (licence: NONE) · get_code("9c625ca32e821ccc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Junghwan-brian/SDE-Net when it was checked. 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. No invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 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 11 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 2026-11-22.

Repositories linked to this paper

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

Abstract

Deep neural networks (NNs) are powerful black box predictors that have recently achieved impressive performance on a wide spectrum of tasks. Quantifying predictive uncertainty in NNs is a challenging and yet unsolved problem. Bayesian NNs, which learn a distribution over weights, are currently the state-of-the-art for estimating predictive uncertainty; however these require significant modifications to the training procedure and are computationally expensive compared to standard (non-Bayesian) NNs. We propose an alternative to Bayesian NNs that is simple to implement, readily parallelizable, requires very little hyperparameter tuning, and yields high quality predictive uncertainty estimates. Through a series of experiments on classification and regression benchmarks, we demonstrate that our method produces well-calibrated uncertainty estimates which are as good or better than approximate Bayesian NNs. To assess robustness to dataset shift, we evaluate the predictive uncertainty on test examples from known and unknown distributions, and show that our method is able to express higher uncertainty on out-of-distribution examples. We demonstrate the scalability of our method by evaluating predictive uncertainty estimates on ImageNet.

For agents

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

get_harvested_code_for_paper("1612.01474")
get_code_for_paper("1612.01474")
have("1612.01474")

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

curl https://syntology.ai/api/ran/1612.01474.json

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

[![Syntology run record](https://syntology.ai/api/ran/1612.01474.svg)](https://syntology.ai/paper/1612.01474)

Connect an agent — have() is free.