SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.09315 · 2022

Deep Linear Networks can Benignly Overfit when Shallow Ones Do

arXiv · PDF · Open in the Atlas

Code that ran

4 of 10 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 4 have no contract verdict; 0 failures of our instrument; 6 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/2209.09315.json · badge: /api/ran/2209.09315.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
niladri-chatterji/benign-deep-linear canonical (classifier label)
classifier basis not recorded on the link
4 ran · 6 no recorded run
FunctionStatusWhere it lives
GaussianIID Ran
ran on a synthesized input
niladri-chatterji/benign-deep-linear/src/data.py
code served (permissive licence) · get_code("0e4bf5add4b1b049")
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 torch

torch.manual_seed(0)

ARGS = {
    "dimension": 4,
    "num_samples": 2,
    "output_dimension": 3,
    "variance": None,
    "k": None,
    "epsilon": None,
    "noise_variance": 1.0,
    "theta": 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 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
TrainNet Ran
ran on a synthesized input
niladri-chatterji/benign-deep-linear/src/utils.py
code served (permissive licence) · get_code("e00b7724c0a7af44")
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 SimpleNet. 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
import torch.optim as optim

torch.manual_seed(0)

# Create a simple neural network
class SimpleNet(nn.Module):
    def __init__(self, input_dim, output_dim):
        super(SimpleNet, self).__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return self.linear(x)

# Create small tensors: batch size 2, input dim 4, output dim 3
X = torch.randn(2, 4)
Y = torch.randn(2, 3)

# Create network
net = SimpleNet(input_dim=4, output_dim=3)

# Create optimizer
optimizer = optim.SGD(net.parameters(), lr=0.01)

# Define ARGS
ARGS = {
    'net': net,
    'X': X,
    'Y': Y,
    'optimizer': optimizer,
    'target_loss': 0.0001,
    'max_epochs': 1000
}
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 SimpleNet — a SimpleNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SimpleNet 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
loss Ran
ran on a synthesized input
niladri-chatterji/benign-deep-linear/src/utils.py
code served (permissive licence) · get_code("3c1e7b0156d6a652")
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 torch

torch.manual_seed(0)

# Create small tensors with distinct sizes
# X: (batch_size, input_dim) = (2, 4)
X = torch.randn(2, 4)

# Y: (batch_size,) or (batch_size, 1) = (2,)
Y = torch.randn(2)

# weights: (input_dim,) = (4,)
weights = torch.randn(4)

ARGS = {
    "X": X,
    "Y": Y,
    "weights": weights
}
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 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
loss_net Ran
ran on a synthesized input
niladri-chatterji/benign-deep-linear/src/utils.py
code served (permissive licence) · get_code("a6ba49d2afc624cf")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create tiny tensors
# X: (batch_size=2, input_dim=4)
X = torch.randn(2, 4)

# Y: (batch_size=2, output_dim=3)
Y = torch.randn(2, 3)

# Create a simple nn.Module that takes input of size 4 and outputs size 3
class SimpleNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

net = SimpleNet()

ARGS = {
    "X": X,
    "Y": Y,
    "net": net
}
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 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
PlantedTwoLayerIID Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/data.py
code served (permissive licence) · get_code("6516763f3d5f80f5")
How it ran

no invocation recorded.

exp1 Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/experiments/exp1.py
code served (permissive licence) · get_code("421fd51a9c0b26f0")
How it ran

no invocation recorded.

exp2 Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/experiments/exp2.py
code served (permissive licence) · get_code("4152f3791094346c")
How it ran

no invocation recorded.

exp3 Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/experiments/exp3.py
code served (permissive licence) · get_code("b61f35aca49e0b22")
How it ran

no invocation recorded.

exp4 Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/experiments/exp4.py
code served (permissive licence) · get_code("fc5fc53e4ba3e925")
How it ran

no invocation recorded.

exp5 Not yet run
unverified (harvested, no recorded run)
niladri-chatterji/benign-deep-linear/src/experiments/exp5.py
code served (permissive licence) · get_code("e7fda969e637db5b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 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 10 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 bound the excess risk of interpolating deep linear networks trained using gradient flow. In a setting previously used to establish risk bounds for the minimum $\ell_2$-norm interpolant, we show that randomly initialized deep linear networks can closely approximate or even match known bounds for the minimum $\ell_2$-norm interpolant. Our analysis also reveals that interpolating deep linear models have exactly the same conditional variance as the minimum $\ell_2$-norm solution. Since the noise affects the excess risk only through the conditional variance, this implies that depth does not improve the algorithm's ability to "hide the noise". Our simulations verify that aspects of our bounds reflect typical behavior for simple data distributions. We also find that similar phenomena are seen in simulations with ReLU networks, although the situation there is more nuanced.

For agents

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

get_harvested_code_for_paper("2209.09315")
get_code_for_paper("2209.09315")
have("2209.09315")

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

curl https://syntology.ai/api/ran/2209.09315.json

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

[![Syntology run record](https://syntology.ai/api/ran/2209.09315.svg)](https://syntology.ai/paper/2209.09315)

Connect an agent — have() is free.