SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2404.11599 · ICLR · 2024

Variational Bayesian Last Layers

Jasper Snoek, James Harrison, John Willes

arXiv · PDF · Open in the Atlas

Code that ran

11 of 23 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 1 violated it and 2 have no contract verdict; 5 failures of our instrument; 12 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/2404.11599.json · badge: /api/ran/2404.11599.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
lightning-uq-box/lightning-uq-box community
Papers with Code archive
6 ran · 6 no recorded run
vectorinstitute/vbll marked official in Papers with Code
Papers with Code archive
5 ran · 6 no recorded run
FunctionStatusWhere it lives
DeterministicRegression Ran
ran on a synthesized input · constructed an object; no result was computed
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("c94c9015de41ea50")
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, DeterministicRegression: 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
from torch.optim import Adam
from lightning.pytorch.cli import LRSchedulerCallable, OptimizerCallable

torch.manual_seed(0)

class SimpleModel(nn.Module):
    """Minimal model for testing."""
    def __init__(self, input_dim: int = 4, output_dim: int = 1):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x: torch.Tensor) -> torch.Tensor:
        return self.linear(x)

class SimpleLoss(nn.Module):
    """Minimal loss function for testing."""
    def forward(self, pred: torch.Tensor, target: torch.Tensor) -> torch.Tensor:
        return nn.functional.mse_loss(pred, target)

model = SimpleModel(input_dim=4, output_dim=1)
loss_fn = SimpleLoss()

ARGS = {
    'model': model,
    'loss_fn': loss_fn,
    'freeze_backbone': False,
    'optimizer': Adam,
    'lr_scheduler': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
lightning, numpy, omegaconf, pandas, torch, torchmetrics — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DeterministicRegression — 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 DeterministicRegression 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
Normal Ran
ran on a synthesized input
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("d7de14910f5c0aef")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 1, dtype family float, shape (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 torch

# The Normal class is a distribution with diagonal covariance
# __init__ takes loc (mean) and chol (scale/std)
# The main computation happens through properties and methods

# Create small tensors for testing
loc = torch.randn(2, 4)  # batch_size=2, feature_dim=4
scale = torch.abs(torch.randn(2, 4)) + 0.1  # ensure positive

INIT_ARGS = {
    'loc': loc,
    'chol': scale,
}

# For testing, we'll call a method that uses the properties
# Let's use covariance_weighted_inner_prod which exercises the class
b = torch.randn(2, 4, 1)  # shape (batch, features, 1)

ARGS = {
    'b': b,
    'reduce_dim': True,
}

METHOD = 'covariance_weighted_inner_prod'
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 1, dtype family float, shape (2,) — shape (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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
_get_input_layer_name_and_module Ran
ran; our contract draft was wrong, not the code
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("b5a8a70192307683")
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 torch
from torch import nn

torch.manual_seed(0)

# Create a simple nested model to test the function
class SimpleInputLayer(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

class MiddleLayer(nn.Module):
    def __init__(self):
        super().__init__()
        self.input_layer = SimpleInputLayer()
    
    def forward(self, x):
        return self.input_layer(x)

class TestModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.first_layer = MiddleLayer()
        self.output = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.first_layer(x)
        x = self.output(x)
        return x

model = TestModel()

ARGS = {
    "model": model
}
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
_get_num_inputs Ran
ran, honoured the contract we drafted
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("916148f6022dcca9")
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 int, ndim 0, dtype family int, 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)

# Build a simple model with a Linear input layer
# The function traverses to the first leaf module and checks for in_features or in_channels
model = nn.Sequential(
    nn.Linear(in_features=4, out_features=8),
    nn.ReLU(),
    nn.Linear(in_features=8, out_features=2)
)

ARGS = {
    'model': model
}
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 int, ndim 0, dtype family int, 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
_get_num_outputs Ran
ran, honoured the contract we drafted
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("ce7423706741a98c")
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 int, ndim 0, dtype family int, 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)

# Create a simple model with an output layer that has out_features
model = nn.Sequential(
    nn.Linear(4, 8),
    nn.ReLU(),
    nn.Linear(8, 3)
)

ARGS = {
    'model': model
}
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 int, ndim 0, dtype family int, 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
_get_output_layer_name_and_module Ran
ran; our contract draft was wrong, not the code
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("1950974c63e5891e")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create a simple model with Linear layers that have out_features
model = nn.Sequential(
    nn.Linear(4, 8),
    nn.ReLU(),
    nn.Linear(8, 3)
)

ARGS = {
    "model": model
}
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
cholesky_inverse Ran
ran; our contract draft was wrong, not the code
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("e2039b4b3c96ebf0")
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 Tensor, ndim 3, dtype family float, shape (2, 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 torch

torch.manual_seed(0)

# Create a small positive definite matrix and compute its Cholesky decomposition
# We'll create a batch of 2 matrices, each 4x4
batch_size = 2
n = 4

# Create positive definite matrices by A = L @ L.T where L is lower triangular
L = torch.randn(batch_size, n, n)
# Make it lower triangular
L = torch.tril(L)
# Ensure positive diagonal for positive definiteness
L = L + torch.eye(n).unsqueeze(0) * 2

# Compute A = L @ L.T to get a positive definite matrix
A = L @ L.transpose(-2, -1)

# Compute Cholesky decomposition (lower triangular)
u_lower = torch.linalg.cholesky(A, upper=False)

# For upper triangular case, transpose
u_upper = u_lower.transpose(-2, -1)

ARGS = {
    "u": u_lower,
    "upper": False,
}
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 Tensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 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
cholupdate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("361e4fe66dd7affa")
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 Tensor, ndim 3, dtype family float, shape (2, 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 torch
import numpy as np

torch.manual_seed(0)

# Create a small positive definite matrix for L (lower triangular Cholesky factor)
matrix_dim = 4
batch_size = 2

# Create a simple lower triangular matrix that is a valid Cholesky factor
# Start with a random matrix and make it lower triangular
L_base = torch.randn(batch_size, matrix_dim, matrix_dim)
L = torch.tril(L_base)
# Ensure diagonal is positive
L = L + torch.eye(matrix_dim).unsqueeze(0) * 2

# Create update vector x with shape (batch_size, matrix_dim)
x = torch.randn(batch_size, matrix_dim)

# weight can be None, a float, or a tensor
weight = 1.0

ARGS = {
    "L": L,
    "x": x,
    "weight": weight,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 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
default_regression_metrics Ran
ran; our contract draft was wrong, not the code
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("1a7a192f2b52410f")
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 MetricCollection. 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 torchmetrics import MetricCollection, MeanSquaredError, MeanAbsoluteError, R2Score

torch.manual_seed(0)

ARGS = {
    'prefix': 'test_'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchmetrics — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MetricCollection — a MetricCollection output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MetricCollection output the recorded ndim 1 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_kl Ran
ran, honoured the contract we drafted
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("bc9b7937ef845522")
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 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

class GaussianDistribution:
    def __init__(self, mean, trace_covariance, logdet_covariance):
        self.mean = mean
        self.trace_covariance = trace_covariance
        self.logdet_covariance = logdet_covariance

# Create a small batch of Gaussian distributions
batch_size = 2
feat_dim = 4

# mean: shape (batch_size, feat_dim)
mean = np.random.randn(batch_size, feat_dim)

# trace_covariance: shape (batch_size,)
trace_covariance = np.abs(np.random.randn(batch_size)) + 0.1

# logdet_covariance: shape (batch_size,)
logdet_covariance = np.random.randn(batch_size)

p = GaussianDistribution(
    mean=mean,
    trace_covariance=trace_covariance,
    logdet_covariance=logdet_covariance
)

q_scale = 2.0

ARGS = {
    "p": p,
    "q_scale": q_scale
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — 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
tp Ran
ran, violated the contract we drafted
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("90e3b209489bfabc")
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 Tensor, ndim 3, dtype family float, shape (2, 3, 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

torch.manual_seed(0)

# The function tp(M) calls M.transpose(-1, -2)
# M must be a tensor with at least 2 dimensions
# Create a small tensor with distinct dimensions for probing
M = torch.randn(2, 4, 3)

ARGS = {
    "M": M
}
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 Tensor, ndim 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
BaseModule Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("d36cc2dac0856730")
How it ran

no invocation recorded.

DenseNormal Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("86d3fa6cc49ee0ca")
How it ran

no invocation recorded.

DenseNormalPrec Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("1c2ac721aed5e052")
How it ran

no invocation recorded.

DeterministicModel Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("071559925536ba24")
How it ran

no invocation recorded.

DiscClassification Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("b25de885c1679905")
How it ran

no invocation recorded.

LowRankNormal Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("4e88d90f40e87668")
How it ran

no invocation recorded.

VBLLRegression Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("e06ace1bf63953c9")
How it ran

no invocation recorded.

VBLLReturn Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("8eba3a2ce2bb8bbd")
How it ran

no invocation recorded.

freeze_model_backbone Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("96779ccf1b2db4f0")
How it ran

no invocation recorded.

get_parameterization Not yet run
unverified (harvested, no recorded run)
vectorinstitute/vbll/vbll/layers/classification.py
code served (permissive licence) · get_code("8916ef4af72a5e59")
How it ran

no invocation recorded.

replace_module Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("26318e664816eac6")
How it ran

no invocation recorded.

save_regression_predictions Not yet run
unverified (harvested, no recorded run)
lightning-uq-box/lightning-uq-box/lightning_uq_box/uq_methods/vbll.py
code served (permissive licence) · get_code("834d1619b99bf13b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 7 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 0 later) and 16 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
BaseModule
brick of this paper's assembly: helper
2 other papers
2 dated earlier · 1 only as an assembly brick
2009.14193 (2020-09)
Uncertainty Sets for Image Classifiers using Conformal Prediction
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
_get_input_layer_name_and_module
brick of this paper's assembly: helper
2 other papers
2 dated earlier · 2 only as an assembly brick
2009.14193 (2020-09)
Uncertainty Sets for Image Classifiers using Conformal Prediction
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
_get_num_inputs
brick of this paper's assembly: helper
2 other papers
2 dated earlier · 2 only as an assembly brick
2009.14193 (2020-09)
Uncertainty Sets for Image Classifiers using Conformal Prediction
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
_get_num_outputs
brick of this paper's assembly: helper
2 other papers
2 dated earlier · 2 only as an assembly brick
2009.14193 (2020-09)
Uncertainty Sets for Image Classifiers using Conformal Prediction
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
_get_output_layer_name_and_module
brick of this paper's assembly: helper
2 other papers
2 dated earlier · 2 only as an assembly brick
2009.14193 (2020-09)
Uncertainty Sets for Image Classifiers using Conformal Prediction
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
DeterministicModel
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference
freeze_model_backbone
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2211.11435 (2022-11)
ZigZag: Universal Sampling-free Uncertainty Estimation Through Two-Step Inference

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

We introduce a deterministic variational formulation for training Bayesian last layer neural networks. This yields a sampling-free, single-pass model and loss that effectively improves uncertainty estimation. Our variational Bayesian last layer (VBLL) can be trained and evaluated with only quadratic complexity in last layer width, and is thus (nearly) computationally free to add to standard architectures. We experimentally investigate VBLLs, and show that they improve predictive accuracy, calibration, and out of distribution detection over baselines across both regression and classification. Finally, we investigate combining VBLL layers with variational Bayesian feature learning, yielding a lower variance collapsed variational inference method for Bayesian neural networks.

For agents

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

get_harvested_code_for_paper("2404.11599")
get_code_for_paper("2404.11599")
have("2404.11599")

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

curl https://syntology.ai/api/ran/2404.11599.json

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

[![Syntology run record](https://syntology.ai/api/ran/2404.11599.svg)](https://syntology.ai/paper/2404.11599)

Connect an agent — have() is free.