SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.05576 · 2020

Self-supervised Learning from a Multi-view Perspective

arXiv · PDF · Open in the Atlas

Code that ran

11 of 15 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 8 have no contract verdict; 3 failures of our instrument; 4 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/2006.05576.json · badge: /api/ran/2006.05576.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
yaohungt/Demystifying_Self_Supervised_Learning canonical (classifier label)
classifier basis: census
11 ran · 4 no recorded run
FunctionStatusWhere it lives
AE_loss Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/Omniglot/compute_MI_CondEntro.py
code served (permissive licence) · get_code("0a7d9a94b7082769")
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 as nn
import torch.nn.functional as F

torch.manual_seed(0)

class SimpleEncoder(nn.Module):
    def __init__(self, input_dim, latent_dim):
        super().__init__()
        self.linear = nn.Linear(input_dim, latent_dim)
    
    def forward(self, x):
        return self.linear(x)

class SimpleDecoder(nn.Module):
    def __init__(self, latent_dim, output_dim):
        super().__init__()
        self.linear = nn.Linear(latent_dim, output_dim)
    
    def forward(self, z):
        return self.linear(z)

# Create encoder and decoder
input_dim = 4
latent_dim = 3
encoder = SimpleEncoder(input_dim, latent_dim)
decoder = SimpleDecoder(latent_dim, input_dim)

# Create input tensors: batch_size=2, features=4
x = torch.randn(2, 4)
s = torch.randn(2, 4)

ARGS = {
    'x': x,
    's': s,
    'encoder': encoder,
    'decoder': decoder
}
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
add_inv_pred_loss Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/objective.py
code served (permissive licence) · get_code("2e338c4bb3d62d7e")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

# Create a small representation tensor of shape (bsz=2, dim=4)
# The function splits this along axis 0, so we need even batch size
repres = tf.constant([
    [1.0, 2.0, 3.0, 4.0],
    [1.5, 2.5, 3.5, 4.5]
], dtype=tf.float32)

# inv_pred_coeff is a scalar coefficient
inv_pred_coeff = 0.5

# weights can be a scalar or compatible shape for weighting the loss
weights = 1.0

ARGS = {
    "repres": repres,
    "inv_pred_coeff": inv_pred_coeff,
    "weights": weights
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
add_supervised_loss Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/objective.py
code served (permissive licence) · get_code("8af989f02c354e14")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create small tensors for the function
batch_size = 2
num_classes = 4

# labels: one-hot encoded labels, shape [batch_size, num_classes]
labels = tf.constant(
    [[1.0, 0.0, 0.0, 0.0],
     [0.0, 1.0, 0.0, 0.0]], 
    dtype=tf.float32
)

# logits: raw model outputs, shape [batch_size, num_classes]
logits = tf.constant(
    [[2.5, 1.2, 0.3, 0.1],
     [0.1, 3.2, 0.5, 0.2]], 
    dtype=tf.float32
)

# weights: per-sample weights, shape [batch_size]
weights = tf.constant([1.0, 1.0], dtype=tf.float32)

ARGS = {
    "labels": labels,
    "logits": logits,
    "weights": weights,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
color_jitter Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/data_util.py
code served (permissive licence) · get_code("324c58e9994efed0")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a small test image tensor
# Image shape: (height, width, channels) with values in [0, 1]
# Using shape (4, 3, 3) for height=4, width=3, channels=3 (RGB)
image_np = np.random.rand(4, 3, 3).astype(np.float32)
image_tensor = tf.constant(image_np)

# Parameters for color_jitter
ARGS = {
    "image": image_tensor,
    "strength": 0.5,
    "random_order": True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
dropblock Ran
ran; our contract draft was wrong, not the code
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/resnet.py
code served (permissive licence) · get_code("7ce61f1a25809b6d")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a small input tensor for the function
# Using channels_first format: [batch, channels, height, width]
# batch=2, channels=3, height=4, width=4
batch_size = 2
channels = 3
spatial_size = 4

net = tf.constant(
    np.random.randn(batch_size, channels, spatial_size, spatial_size),
    dtype=tf.float32
)

is_training = True
keep_prob = 0.9
dropblock_size = 2
data_format = 'channels_first'

ARGS = {
    'net': net,
    'is_training': is_training,
    'keep_prob': keep_prob,
    'dropblock_size': dropblock_size,
    'data_format': data_format
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
fixed_padding Ran
ran; our contract draft was wrong, not the code
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/resnet.py
code served (permissive licence) · get_code("3223825e26df85e7")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a small tensor in channels_first format: [batch, channels, height, width]
batch_size = 2
channels = 3
height = 4
width = 5

# Test with channels_first format
inputs_channels_first = tf.constant(
    np.random.randn(batch_size, channels, height, width).astype(np.float32)
)

kernel_size = 3
data_format = 'channels_first'

ARGS = {
    'inputs': inputs_channels_first,
    'kernel_size': kernel_size,
    'data_format': data_format
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
init_models Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/Omniglot/compute_MI_CondEntro.py
code served (permissive licence) · get_code("1eed2bc576d1a00a")
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 1, dtype family float; no output shape is stated. 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 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 torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'stage': 'AE',
    'additional_encoder': False
}
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, ndim 1, 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
mlp Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/Omniglot/compute_MI_CondEntro.py
code served (permissive licence) · get_code("b509a54eb5ec250b")
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. 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)

ARGS = {
    "dim": 4,
    "hidden_dim": 8,
    "output_dim": 2,
    "layers": 2,
    "batch_norm": False,
}
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
pad_to_batch Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/data.py
code served (permissive licence) · get_code("0c26c80b6286f5ec")
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 DatasetV1Adapter. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a simple dataset with tensors of shape (2, 4)
# The dataset should yield individual elements that will be batched
def create_dataset():
    # Create sample data: 3 samples, each with shape (4,)
    data = 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]], dtype=np.float32)
    
    # Create a tf.data.Dataset from the numpy array
    dataset = tf.data.Dataset.from_tensor_slices(data)
    return dataset

dataset = create_dataset()
batch_size = 4

ARGS = {
    'dataset': dataset,
    'batch_size': batch_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DatasetV1Adapter — a DatasetV1Adapter output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DatasetV1Adapter 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
random_apply Ran
ran on a synthesized input
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/data_util.py
code served (permissive licence) · get_code("d04f187058cb2a9c")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

def simple_transform(x):
    """Simple transformation function that doubles the input."""
    return x * 2

# Create a small tensor
x = tf.constant([1.0, 2.0, 3.0, 4.0], dtype=tf.float32)

# Define probability
p = 0.7

# Define function to apply
func = simple_transform

ARGS = {
    "func": func,
    "p": p,
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
tpu_cross_replica_concat Ran
ran; our contract draft was wrong, not the code
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/objective.py
code served (permissive licence) · get_code("7aac0de5d1568953")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

class TPUContext:
    """Stand-in TPUContext class with required attributes."""
    def __init__(self, num_replicas=1):
        self.num_replicas = num_replicas

# Create a small tensor (batch=2, features=4)
tensor = tf.constant([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32)

# Create a TPUContext with num_replicas > 1 to exercise the main logic
tpu_context = TPUContext(num_replicas=2)

ARGS = {
    "tensor": tensor,
    "tpu_context": tpu_context,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
get_train_steps Not yet run
unverified (harvested, no recorded run)
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/model_util.py
code served (permissive licence) · get_code("14ec8ff095316493")
How it ran

no invocation recorded.

learning_rate_schedule Not yet run
unverified (harvested, no recorded run)
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/model_util.py
code served (permissive licence) · get_code("fb45bd9f63767265")
How it ran

no invocation recorded.

to_grayscale Not yet run
unverified (harvested, no recorded run)
yaohungt/Demystifying_Self_Supervised_Learning/CIFAR10/data_util.py
code served (permissive licence) · get_code("49c827442d2fb761")
How it ran

no invocation recorded.

train_val Not yet run
unverified (harvested, no recorded run)
yaohungt/Demystifying_Self_Supervised_Learning/Omniglot/linear.py
pointer only (licence: MIT); the stored copy get_code serves, from leftthomas/SimCLR, is pointer only (licence: NONE) · get_code("667eaf3785a1c4a9")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from leftthomas/SimCLR, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 10 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 10 later) and 5 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.

FunctionAlso present inEarliest-datedOthers named
get_train_steps 4 other papers
1 the same month, 3 later
2006.10029 (2020-06)
Big Self-Supervised Models are Strong Semi-Supervised Learners
2109.12909 (2021-09)
Compressive Visual Representations
aaai_20650 (2022 (year only))
Adversarial Examples Can Be Effective Data Augmentation for Unsupervised Machine Learning
and 1 more not named here
random_apply 4 other papers
4 later
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2108.10310 (2021-08)
Ranking Models in Unlabeled New Environments
2112.11435 (2021-12)
Learned Queries for Efficient Local Attention
and 1 more not named here
to_grayscale 4 other papers
4 later
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2108.10310 (2021-08)
Ranking Models in Unlabeled New Environments
2109.12909 (2021-09)
Compressive Visual Representations
and 1 more not named here
fixed_padding 3 other papers
2 dated earlier, 1 later
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
1907.10701 (2019-07)
Benchmarking TPU, GPU, and CPU Platforms for Deep Learning
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision
tpu_cross_replica_concat 2 other papers
1 dated earlier, 1 later
2002.05709 (2020-02)
A Simple Framework for Contrastive Learning of Visual Representations
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision
train_val 2 other papers
1 dated earlier, 1 later
2002.05709 (2020-02)
A Simple Framework for Contrastive Learning of Visual Representations
2007.00224 (2020-07)
Debiased Contrastive Learning
add_supervised_loss 1 other paper
1 later
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision
dropblock 1 other paper
1 later
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision
learning_rate_schedule 1 other paper
1 later
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision
pad_to_batch 1 other paper
1 later
2205.09723 (2022-05)
Robust and Efficient Medical Imaging with Self-Supervision

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

As a subset of unsupervised representation learning, self-supervised representation learning adopts self-defined signals as supervision and uses the learned representation for downstream tasks, such as object detection and image captioning. Many proposed approaches for self-supervised learning follow naturally a multi-view perspective, where the input (e.g., original images) and the self-supervised signals (e.g., augmented images) can be seen as two redundant views of the data. Building from this multi-view perspective, this paper provides an information-theoretical framework to better understand the properties that encourage successful self-supervised learning. Specifically, we demonstrate that self-supervised learned representations can extract task-relevant information and discard task-irrelevant information. Our theoretical framework paves the way to a larger space of self-supervised learning objective design. In particular, we propose a composite objective that bridges the gap between prior contrastive and predictive learning objectives, and introduce an additional objective term to discard task-irrelevant information. To verify our analysis, we conduct controlled experiments to evaluate the impact of the composite objectives. We also explore our framework's empirical generalization beyond the multi-view perspective, where the cross-view redundancy may not be clearly observed.

For agents

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

get_harvested_code_for_paper("2006.05576")
get_code_for_paper("2006.05576")
have("2006.05576")

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

curl https://syntology.ai/api/ran/2006.05576.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.05576.svg)](https://syntology.ai/paper/2006.05576)

Connect an agent — have() is free.