SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2011.00515 · ICML · 2021

On Signal-to-Noise Ratio Issues in Variational Inference for Deep Gaussian Processes

Yarin Gal, Tom Rainforth, Tim Rudner, Oscar Key

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 2 failures of our instrument; 0 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/2011.00515.json · badge: /api/ran/2011.00515.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
timrudner/snr_issues_in_deep_gps canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
get_k Ran
ran, honoured the contract we drafted
timrudner/snr_issues_in_deep_gps/experiments/create_figures.py
code served (permissive licence) · get_code("49ab52bcc1f72492")
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 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 os

def get_k(file_name) -> int:
    parts = os.path.splitext(file_name)[0].split('_')
    for part in parts:
        if part.startswith('k'):
            return int(part.replace('k', ''))
    raise ValueError(f'missing k def: {file_name}')

ARGS = {
    'file_name': 'model_k5_checkpoint.pt'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — 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
group_by_M Ran
ran; our contract draft was wrong, not the code
timrudner/snr_issues_in_deep_gps/experiments/create_figures.py
code served (permissive licence) · get_code("c9c49c3ed975bca6")
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 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

# Test case 1: 1D gradient samples
gradient_samples_1d = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0])
M_1d = 2

# Test case 2: 2D gradient samples
gradient_samples_2d = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0],
    [9.0, 10.0, 11.0, 12.0],
    [13.0, 14.0, 15.0, 16.0],
    [17.0, 18.0, 19.0, 20.0],
    [21.0, 22.0, 23.0, 24.0]
])
M_2d = 3

# Use 1D case for ARGS
ARGS = {
    "gradient_samples": gradient_samples_1d,
    "M": M_1d
}
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 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
load_gradients Ran
ran; our contract draft was wrong, not the code
timrudner/snr_issues_in_deep_gps/experiments/create_figures.py
code served (permissive licence) · get_code("173c6baf6062e84d")
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 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 os
import tempfile
import numpy as np
from pathlib import Path

# Create a temporary directory structure for the test
temp_dir = tempfile.mkdtemp()
checkpoints_dir = os.path.join(temp_dir, 'checkpoints', 'test_experiment')
os.makedirs(checkpoints_dir, exist_ok=True)

# Create sample gradient files with different k values
# Format: gradients_bs{batch_size}{dreg_ext}_k{K}.npy
for k_val in [10, 50, 100, 500]:
    # Create a dict with gradient data
    # The function expects a dict with parameter names as keys and gradient arrays as values
    gradients_dict = {
        'W_layer1': np.random.randn(4, 3),  # Shape (N, features) where N=4
        'W_layer2': np.random.randn(4, 5),  # Shape (N, features) where N=4
        'b_layer1': np.random.randn(4, 2),  # Will be flattened
    }
    
    file_path = os.path.join(checkpoints_dir, f'gradients_bs1_k{k_val}.npy')
    np.save(file_path, gradients_dict)

# Also create a dreg version for testing
for k_val in [10, 50]:
    gradients_dict = {
        'W_layer1': np.random.randn(4, 3),
        'W_layer2': np.random.randn(4, 5),
    }
    
    file_path = os.path.join(checkpoints_dir, f'gradients_bs1_dreg_k{k_val}.npy')
    np.save(file_path, gradients_dict)

ARGS = {
    'results_dir': temp_dir,
    'experiment_name': 'test_experiment',
    'dreg': False,
    'batch_size': 1,
    'max_k': 100000,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, tqdm — 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

We show that the gradient estimates used in training Deep Gaussian Processes (DGPs) with importance-weighted variational inference are susceptible to signal-to-noise ratio (SNR) issues. Specifically, we show both theoretically and via an extensive empirical evaluation that the SNR of the gradient estimates for the latent variable's variational parameters decreases as the number of importance samples increases. As a result, these gradient estimates degrade to pure noise if the number of importance samples is too large. To address this pathology, we show how doubly reparameterized gradient estimators, originally proposed for training variational autoencoders, can be adapted to the DGP setting and that the resultant estimators completely remedy the SNR issue, thereby providing more reliable training. Finally, we demonstrate that our fix can lead to consistent improvements in the predictive performance of DGP models.

For agents

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

get_harvested_code_for_paper("2011.00515")
get_code_for_paper("2011.00515")
have("2011.00515")

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

curl https://syntology.ai/api/ran/2011.00515.json

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

[![Syntology run record](https://syntology.ai/api/ran/2011.00515.svg)](https://syntology.ai/paper/2011.00515)

Connect an agent — have() is free.