SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.04994 · ICLR · 2023

Sampling-based inference for large linear models, with application to linearised Laplace

Javier Antorán, José Hernández-Lobato, Eric Nalisnick, David Janz, Riccardo Barbano, Shreyas Padhy

arXiv · PDF · Open in the Atlas

Code that ran

2 of 4 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 2 have no contract verdict; 0 failures of our instrument; 2 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/2210.04994.json · badge: /api/ran/2210.04994.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
cambridge-mlg/sampled-laplace canonical (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
FunctionStatusWhere it lives
hessian_of_loss_fn Ran
ran on a synthesized input
cambridge-mlg/sampled-laplace/src/utils.py
code served (permissive licence) · get_code("b20e76f089377b5c")
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 ArrayImpl, 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 jax
import jax.numpy as jnp
import chex

# Set random seed for reproducibility
jax.random.PRNGKey(0)

# Create a small logits array for a classification problem
# Shape: (num_classes,) - single point prediction
# Using 4 classes as an example
logits = jnp.array([1.0, 2.0, 0.5, -0.5], dtype=jnp.float32)

# Optional jitter parameter for numerical stability
H_L_jitter = 1e-6

# Loss function type
loss = 'crossentropy'

ARGS = {
    'logits': logits,
    'H_L_jitter': H_L_jitter,
    'loss': loss,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
chex, jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
tf_target_samples_dataset Ran
ran on a synthesized input
cambridge-mlg/sampled-laplace/src/data.py
code served (permissive licence) · get_code("b3fe088260072037")
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 _ParallelMapDataset. 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 h5py
import numpy as np
import tensorflow as tf
import tempfile
import os

# Create a temporary h5 file with target samples
temp_dir = tempfile.mkdtemp()
target_samples_path = os.path.join(temp_dir, 'target_samples.h5')

# Create h5 file with required structure
with h5py.File(target_samples_path, 'w') as f:
    # Create target_samples: shape (2, 3, 5) - 2 samples, 3x3 spatial, 5 channels
    target_samples = np.random.randn(2, 3, 5).astype(np.float32)
    f.create_dataset('target_samples', data=target_samples)
    
    # Create targets (labels): shape (2,) - 2 labels
    targets = np.array([0, 1], dtype=np.int64)
    f.create_dataset('targets', data=targets)

# Create image_dataset: tf.data.Dataset that yields dicts with 'image' and 'label'
def create_image_dataset():
    images = np.random.randn(2, 4, 4, 3).astype(np.float32)  # 2 images, 4x4, 3 channels
    labels = np.array([0, 1], dtype=np.int64)
    
    def gen():
        for img, label in zip(images, labels):
            yield {'image': img, 'label': label}
    
    output_signature = {
        'image': tf.TensorSpec(shape=(4, 4, 3), dtype=tf.float32),
        'label': tf.TensorSpec(shape=(), dtype=tf.int64)
    }
    dataset = tf.data.Dataset.from_generator(gen, output_signature=output_signature)
    return dataset

image_dataset = create_image_dataset()

ARGS = {
    'image_dataset': image_dataset,
    'target_samples_path': target_samples_path,
    'num_samples': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
h5py, numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type _ParallelMapDataset — a _ParallelMapDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _ParallelMapDataset 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
compute_hessian_of_loss Not yet run
unverified (harvested, no recorded run)
cambridge-mlg/sampled-laplace/src/utils.py
code served (permissive licence) · get_code("adda74776b1dc4e2")
How it ran

no invocation recorded.

compute_hessian_of_loss_batched Not yet run
unverified (harvested, no recorded run)
cambridge-mlg/sampled-laplace/src/utils.py
code served (permissive licence) · get_code("47770d6b9137b068")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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

Large-scale linear models are ubiquitous throughout machine learning, with contemporary application as surrogate models for neural network uncertainty quantification; that is, the linearised Laplace method. Alas, the computational cost associated with Bayesian linear models constrains this method's application to small networks, small output spaces and small datasets. We address this limitation by introducing a scalable sample-based Bayesian inference method for conjugate Gaussian multi-output linear models, together with a matching method for hyperparameter (regularisation strength) selection. Furthermore, we use a classic feature normalisation method, the g-prior, to resolve a previously highlighted pathology of the linearised Laplace method. Together, these contributions allow us to perform linearised neural network inference with ResNet-18 on CIFAR100 (11M parameters, 100 outputs × 50k datapoints), with ResNet-50 on Imagenet (50M parameters, 1000 outputs × 1.2M datapoints) and with a U-Net on a high-resolution tomographic reconstruction task (2M parameters, 251k output dimensions).

For agents

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

get_harvested_code_for_paper("2210.04994")
get_code_for_paper("2210.04994")
have("2210.04994")

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

curl https://syntology.ai/api/ran/2210.04994.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.04994.svg)](https://syntology.ai/paper/2210.04994)

Connect an agent — have() is free.