SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1806.05413 · 2018

Learning Dynamics of Linear Denoising Autoencoders

arXiv · PDF · Open in the Atlas

Code that ran

14 of 19 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 14 have no contract verdict; 0 failures of our instrument; 5 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/1806.05413.json · badge: /api/ran/1806.05413.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
arnupretorius/lindaedynamics_icml2018 canonical (classifier label)
classifier basis: census
14 ran · 5 no recorded run
FunctionStatusWhere it lives
compute_correlation_matrix Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/linear_ae_net/dynamics.py
code served (permissive licence) · get_code("94a23729745a0c4d")
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, 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 small numpy arrays with distinct shapes
# x will be (batch_size=2, x_dim=4)
# y will be (batch_size=2, y_dim=3)
x = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float32)

y = np.array([
    [0.1, 0.2, 0.3],
    [0.4, 0.5, 0.6]
], dtype=np.float32)

ARGS = {
    "x": x,
    "y": y
}
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, 4) — shape (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
create_model Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/nonlinear_ae_net/nonlinear_ae_net.py
code served (permissive licence) · get_code("09cd88e02a7b5179")
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 Sequential. 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)

ARGS = {
    'D_in': 4,
    'H': 8,
    'D_out': 3,
    'act_fn': 'tanh',
    'noise': None
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential 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
dae_learning_dynamics Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/linear_ae_net/dynamics.py
code served (permissive licence) · get_code("9077a88c2c93afef")
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 (10,) 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 = {
    'lam': 0.5,
    'var': 0.1,
    'gamma': 0.02,
    'N': 100,
    'n_epoch': 10,
    'learning_rate': 0.01,
    'u0': 1.0
}
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 (10,) — shape (10,) 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
generate_circle_points Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/utils.py
code served (permissive licence) · get_code("1936ca8c49dcffd2")
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

# Read the function signature:
# def generate_circle_points(r, n=1):
# 
# Parameters:
# - r: radius (used as a scalar multiplier in np.cos and np.sin)
# - n: number of points (used in range(0, n))

ARGS = {
    "r": 5.0,
    "n": 10
}
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 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
hyperbolic_learning_dynamics Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/utils.py
code served (permissive licence) · get_code("456c9738bf586056")
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 list, 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 numpy as np

# Construct arguments for hyperbolic_learning_dynamics
x = 2.5
y = 1.3
n_epoch = 5
learning_rate = 0.01
w1_init = 0.5
w2_init = 0.7
l2 = 0.1

ARGS = {
    "x": x,
    "y": y,
    "n_epoch": n_epoch,
    "learning_rate": learning_rate,
    "w1_init": w1_init,
    "w2_init": w2_init,
    "l2": l2
}
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 list, ndim 1, dtype family float — a list 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
mean_squared_error_backward Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/loss_functions.py
code served (permissive licence) · get_code("5dd7f0a2a4002782")
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

# Construct cache tuple: (err, y_true)
# err should be the error (y_pred - y_true), typically shape (batch_size, output_dim)
# y_true should be the true labels, same shape as err
err = np.array([[0.5, -0.2, 0.1, 0.3],
                [0.2, 0.4, -0.1, 0.0]], dtype=np.float32)

y_true = np.array([[1.0, 0.0, 1.0, 0.5],
                   [0.0, 1.0, 0.5, 1.0]], dtype=np.float32)

cache = (err, y_true)

# dout is the upstream gradient
dout = np.array([[0.1, 0.2, 0.15, 0.05],
                 [0.3, 0.1, 0.2, 0.25]], dtype=np.float32)

ARGS = {
    "dout": dout,
    "cache": cache
}
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
mean_squared_error_forward Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/loss_functions.py
code served (permissive licence) · get_code("495ac14b8cbc4cfc")
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. 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

# Create small numpy arrays for y_pred and y_true
y_pred = np.array([[0.1, 0.2, 0.3, 0.4],
                   [0.5, 0.6, 0.7, 0.8]])

y_true = np.array([[1.0, 0.0, 1.0, 0.0],
                   [0.0, 1.0, 0.0, 1.0]])

ARGS = {
    'y_pred': y_pred,
    'y_true': y_true
}
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 tuple — a tuple output has no array shape of its own; none is 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_init Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/initialisers.py
code served (permissive licence) · get_code("de719648984aa476")
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 float, ndim 0, dtype family float, 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 numpy as np

def normal_init(weight_scale, loc=0, scale=1):
    w = weight_scale * np.random.normal(loc, scale)
    return w

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

ARGS = {
    "weight_scale": 0.5,
    "loc": 0,
    "scale": 1
}
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 float, ndim 0, dtype family float, 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
optimise Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/optimisers.py
code served (permissive licence) · get_code("8af790e81422aab5")
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 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Define a simple optimizer function that takes params, grads, and learning_rate
def simple_optimizer(param, grad, learning_rate):
    """Simple SGD-like optimizer that updates param by subtracting learning_rate * grad"""
    return param - learning_rate * grad

# Create small tensor parameters
params = {
    'weight': torch.randn(4, 3),
    'bias': torch.randn(3)
}

# Create gradients with same shapes as params
grads = {
    'weight': torch.randn(4, 3),
    'bias': torch.randn(3)
}

# Define the optimizer function
optimiser = simple_optimizer

# Set learning rate
learning_rate = 0.01

ARGS = {
    'params': params,
    'grads': grads,
    'optimiser': optimiser,
    'learning_rate': learning_rate
}
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 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
plot_mse_loss_surface_2d Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/visualisations.py
code served (permissive licence) · get_code("204b7e13e88d206c")
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. 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
import matplotlib.pyplot as plt

# Create figure and axis
fig, ax = plt.subplots()

# Create small sample data
# x and y are used in the loss computation: y_pred = ws_x[i, j] * ws_y[i, j] * x
# and y_true = y, then loss includes (y_true - y_pred) ** 2
# So x and y should be scalar or array-like
x = np.array(2.5)
y = np.array(1.0)

# v and l2 are regularization parameters (scalars)
v = 0.1
l2 = 0.01

# w1_range and w2_range define the weight space ranges to plot
w1_range = (-2, 2)
w2_range = (2, -2)

ARGS = {
    'fig': fig,
    'ax': ax,
    'x': x,
    'y': y,
    'v': v,
    'l2': l2,
    'w1_range': w1_range,
    'w2_range': w2_range,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
relu_forward Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/layers.py
code served (permissive licence) · get_code("4fe3aaa057cb62f9")
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 3, 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

# relu_forward takes a single parameter 'x' which is used with np.maximum
# The function performs element-wise ReLU: max(0, x)
# x should be a numpy array

# Create a small test input array with negative and positive values
x = np.array([[-2.0, 1.0, -0.5, 3.0],
              [0.5, -1.5, 2.0, -0.1]], dtype=np.float32)

ARGS = {
    'x': x
}
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 tuple, ndim 3, 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
theoretical_dynamics Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/utils.py
code served (permissive licence) · get_code("a6b47b98df4d0ce7")
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 list, 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 numpy as np

# Analyze the function signature and body:
# - x: used in arithmetic (2 * x * y * t / tau), appears to be a scalar float
# - y: used in arithmetic (2 * x * y * t / tau, y * E, y / u0), appears to be a scalar float
# - w1_init: multiplied with w2_init to form u0, scalar float
# - w2_init: multiplied with w1_init to form u0, scalar float
# - n_epoch: loop range, integer
# - learning_rate: used to compute tau = 1 / learning_rate, scalar float

ARGS = {
    "x": 0.5,
    "y": 2.0,
    "w1_init": 1.5,
    "w2_init": 0.8,
    "n_epoch": 5,
    "learning_rate": 0.1
}
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 list, ndim 1, dtype family float — a list 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
theoretical_learning_dynamics Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/linear_ae_net/dynamics.py
code served (permissive licence) · get_code("de50c529f4b5188c")
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 (10, 3) 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

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

# Create small sample data
# X: (n_samples, n_features) - using distinct dimensions 5x4
X = np.random.randn(5, 4)

# y: (n_samples, n_outputs) - using distinct dimensions 5x3
y = np.random.randn(5, 3)

# Other parameters
n_epoch = 10
lr = 0.01
var = 0.001
reg = 0.0001
u0 = 2.5e-07

ARGS = {
    "X": X,
    "y": y,
    "n_epoch": n_epoch,
    "lr": lr,
    "var": var,
    "reg": reg,
    "u0": u0
}
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 (10, 3) — shape (10, 3) 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
uniform_init Ran
ran on a synthesized input
arnupretorius/lindaedynamics_icml2018/src/scalar_net/initialisers.py
code served (permissive licence) · get_code("3b15ad22ad03850b")
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 float, ndim 0, dtype family float, 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 numpy as np

np.random.seed(0)

ARGS = {
    "weight_scale": 2.5,
    "a": -1,
    "b": 1,
}
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 float, ndim 0, dtype family float, 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
affine_backward Not yet run
unverified (harvested, no recorded run)
arnupretorius/lindaedynamics_icml2018/src/scalar_net/layers.py
code served (permissive licence) · get_code("14f8157279eda080")
How it ran

no invocation recorded.

affine_forward Not yet run
unverified (harvested, no recorded run)
arnupretorius/lindaedynamics_icml2018/src/scalar_net/layers.py
code served (permissive licence) · get_code("5f1182602605491b")
How it ran

no invocation recorded.

animate_optimiser_trajectory_2d Not yet run
unverified (harvested, no recorded run)
arnupretorius/lindaedynamics_icml2018/src/scalar_net/visualisations.py
code served (permissive licence) · get_code("e019ef114b73880b")
How it ran

no invocation recorded.

gradient_descent Not yet run
unverified (harvested, no recorded run)
arnupretorius/lindaedynamics_icml2018/src/scalar_net/optimisers.py
code served (permissive licence) · get_code("5676f4e445be4c1e")
How it ran

no invocation recorded.

plot_mse_loss_surface_3d Not yet run
unverified (harvested, no recorded run)
arnupretorius/lindaedynamics_icml2018/src/scalar_net/visualisations.py
code served (permissive licence) · get_code("4d3e418f768edb34")
How it ran

no invocation recorded.

Also present in other papers' code

Of 19 samples lifted from this paper's linked repositories, 19 carry a function identity and 0 do not. Of the 19 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 19 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

Denoising autoencoders (DAEs) have proven useful for unsupervised representation learning, but a thorough theoretical understanding is still lacking of how the input noise influences learning. Here we develop theory for how noise influences learning in DAEs. By focusing on linear DAEs, we are able to derive analytic expressions that exactly describe their learning dynamics. We verify our theoretical predictions with simulations as well as experiments on MNIST and CIFAR-10. The theory illustrates how, when tuned correctly, noise allows DAEs to ignore low variance directions in the inputs while learning to reconstruct them. Furthermore, in a comparison of the learning dynamics of DAEs to standard regularised autoencoders, we show that noise has a similar regularisation effect to weight decay, but with faster training dynamics. We also show that our theoretical predictions approximate learning dynamics on real-world data and qualitatively match observed dynamics in nonlinear DAEs.

For agents

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

get_harvested_code_for_paper("1806.05413")
get_code_for_paper("1806.05413")
have("1806.05413")

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

curl https://syntology.ai/api/ran/1806.05413.json

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

[![Syntology run record](https://syntology.ai/api/ran/1806.05413.svg)](https://syntology.ai/paper/1806.05413)

Connect an agent — have() is free.