SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1905.05928 · 2019

Rethinking the Usage of Batch Normalization and Dropout in the Training of Deep Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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 7 have no contract verdict; 0 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/1905.05928.json · badge: /api/ran/1905.05928.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
tae898/age-gender community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran on a synthesized input
tae898/age-gender/model/metric.py
code served (permissive licence) · get_code("2edae59f843d8ce9")
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)

# The function expects:
# - output: a tensor with shape (batch_size, num_classes) containing logits/scores
# - target: a tensor with shape (batch_size,) containing class indices

batch_size = 2
num_classes = 4

# Create output tensor (logits for 2 samples, 4 classes)
output = torch.randn(batch_size, num_classes)

# Create target tensor (class indices for 2 samples)
target = torch.tensor([0, 2])

ARGS = {
    "output": output,
    "target": target,
}
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
accuracy_mse Ran
ran on a synthesized input
tae898/age-gender/model/metric.py
code served (permissive licence) · get_code("38ab19fce92aedaf")
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 test tensors
# output: predicted values from a regression model
# target: ground truth values
# The function compares absolute differences < 1

output = torch.randn(2, 4)  # batch size 2, 4 dimensions
target = torch.randn(2, 4)  # batch size 2, 4 dimensions

ARGS = {
    "output": output,
    "target": target,
}
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
accuracy_relaxed Ran
ran on a synthesized input
tae898/age-gender/model/metric.py
code served (permissive licence) · get_code("7305e7b9b187765b")
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 output tensor with shape (batch_size, num_classes)
# The function calls torch.argmax(output, dim=1) to get predictions
# So output needs to be 2D: (batch_size, num_classes)
output = torch.randn(2, 4)  # batch size 2, 4 classes

# Create target tensor with shape (batch_size,)
# The function zips pred and target, so target must be 1D with same length as pred
target = torch.tensor([1, 5])  # batch size 2, integer class indices

ARGS = {
    "output": output,
    "target": target,
}
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
cse Ran
ran on a synthesized input
tae898/age-gender/model/loss.py
code served (permissive licence) · get_code("bfb99459d150b0f0")
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 Tensor, 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.functional as F

torch.manual_seed(0)

# The function takes two tensors and computes cross entropy loss
# F.cross_entropy expects:
# - output: logits of shape (batch_size, num_classes)
# - target: class indices of shape (batch_size,) with values in [0, num_classes)

batch_size = 2
num_classes = 4

# Create output tensor with logits (batch_size, num_classes)
output = torch.randn(batch_size, num_classes)

# Create target tensor with class indices (batch_size,)
target = torch.tensor([0, 2], dtype=torch.long)

ARGS = {
    "output": output,
    "target": target
}
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 Tensor, 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
mse Ran
ran on a synthesized input
tae898/age-gender/model/loss.py
code served (permissive licence) · get_code("1dc2c8c98c52e761")
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 Tensor, 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.functional as F

torch.manual_seed(0)

# Create small tensors for output and target
# Using distinct dimensions (batch_size=2, features=4)
output = torch.randn(2, 4)
target = torch.randn(2, 4)

ARGS = {
    "output": output,
    "target": target,
}
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 Tensor, 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
nll_loss Ran
ran on a synthesized input
tae898/age-gender/model/loss.py
code served (permissive licence) · get_code("f733c7f26bd8c1dc")
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 Tensor, 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.functional as F

torch.manual_seed(0)

# nll_loss expects:
# - output: a tensor with log probabilities (typically from log_softmax)
# - target: a tensor with class indices

# Create a batch of 2 samples with 4 classes
# output should be log probabilities
output = F.log_softmax(torch.randn(2, 4), dim=1)

# target should be class indices (LongTensor)
target = torch.tensor([0, 3], dtype=torch.long)

ARGS = {
    'output': output,
    'target': target,
}
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 Tensor, 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
send_to_servers Ran
ran on a synthesized input
tae898/age-gender/client.py
code served (permissive licence) · get_code("25e3ac309687175f")
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 logging
import pickle
import jsonpickle
import numpy as np
from unittest.mock import Mock, patch

# Set up logging
logging.basicConfig(level=logging.DEBUG)

# Create a mock binary image (bytes)
binary_image = b'\x89PNG\r\n\x1a\n\x00\x00\x00\rIHDR\x00\x00\x00\x01'

# Create mock server URLs
url_face = "http://localhost:5000/face"
url_age_gender = "http://localhost:5001/age_gender"

# Mock the requests.post responses
def mock_face_detection_response(*args, **kwargs):
    """Mock response from face detection server"""
    face_data = {
        'face_detection_recognition': [
            {
                'bbox': [10, 20, 100, 150],
                'det_score': 0.95,
                'landmark': [[15, 25], [30, 35], [45, 50], [60, 65], [75, 80]],
                'normed_embedding': np.random.randn(512).astype(np.float32).tolist()
            },
            {
                'bbox': [200, 220, 280, 320],
                'det_score': 0.87,
                'landmark': [[205, 225], [220, 235], [235, 250], [250, 265], [265, 280]],
                'normed_embedding': np.random.randn(512).astype(np.float32).tolist()
            }
        ]
    }
    response = Mock()
    response.text = jsonpickle.encode(face_data)
    return response

def mock_age_gender_response(*args, **kwargs):
    """Mock response from age-gender server"""
    age_gender_data = {
        'ages': [28, 35],
        'genders': ['M', 'F']
    }
    response = Mock()
    response.text = jsonpickle.encode(age_gender_data)
    return response

# Patch requests.post to return mock responses
import requests
original_post = requests.post

def patched_post(url, *args, **kwargs):
    if 'face' in url:
        return mock_face_detection_response(*args, **kwargs)
    elif 'age_gender' in url:
        return mock_age_gender_response(*args, **kwargs)
    return original_post(url, *args, **kwargs)

requests.post = patched_post

ARGS = {
    'binary_image': binary_image,
    'url_face': url_face,
    'url_age_gender': url_age_gender
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jsonpickle, numpy, requests — 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

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 7 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

In this work, we propose a novel technique to boost training efficiency of a neural network. Our work is based on an excellent idea that whitening the inputs of neural networks can achieve a fast convergence speed. Given the well-known fact that independent components must be whitened, we introduce a novel Independent-Component (IC) layer before each weight layer, whose inputs would be made more independent. However, determining independent components is a computationally intensive task. To overcome this challenge, we propose to implement an IC layer by combining two popular techniques, Batch Normalization and Dropout, in a new manner that we can rigorously prove that Dropout can quadratically reduce the mutual information and linearly reduce the correlation between any pair of neurons with respect to the dropout layer parameter $p$. As demonstrated experimentally, the IC layer consistently outperforms the baseline approaches with more stable training process, faster convergence speed and better convergence limit on CIFAR10/100 and ILSVRC2012 datasets. The implementation of our IC layer makes us rethink the common practices in the design of neural networks. For example, we should not place Batch Normalization before ReLU since the non-negative responses of ReLU will make the weight layer updated in a suboptimal way, and we can achieve better performance by combining Batch Normalization and Dropout together as an IC layer.

For agents

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

get_harvested_code_for_paper("1905.05928")
get_code_for_paper("1905.05928")
have("1905.05928")

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

curl https://syntology.ai/api/ran/1905.05928.json

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

[![Syntology run record](https://syntology.ai/api/ran/1905.05928.svg)](https://syntology.ai/paper/1905.05928)

Connect an agent — have() is free.