SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_32174 · AAAI · 2025

Can Generative Models Improve Self-Supervised Representation Learning?

Arash Afkanpour, Sana Ayromlou, Vahid Khazaie, Fereshteh Forghani

Open in the Atlas

Code that ran

10 of 12 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 0 have no contract verdict; 10 failures of our instrument; 2 no recorded run), measured 2026-11-22; 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/aaai_32174.json · badge: /api/ran/aaai_32174.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
VectorInstitute/GenerativeSSL canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
FunctionStatusWhere it lives
barlow_loss_func Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/solo-learn/solo/losses/barlow.py
code served (permissive licence) · get_code("132a949a804eb81e")
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

torch.manual_seed(0)

# Create small tensors for z1 and z2
# N=2 (batch size), D=4 (feature dimension)
z1 = torch.randn(2, 4)
z2 = torch.randn(2, 4)

# Set standard hyperparameters
lamb = 0.005
scale_loss = 0.025

ARGS = {
    "z1": z1,
    "z2": z2,
    "lamb": lamb,
    "scale_loss": scale_loss,
}
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
byol_loss_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
VectorInstitute/GenerativeSSL/solo-learn/solo/losses/byol.py
pointer only (licence: MIT); the stored copy get_code serves, from PKU-ML/nlb, is pointer only (licence: NONE) · get_code("273532e06cb3967e")
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 PKU-ML/nlb, 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. A run is recorded.

gram_schmidt Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/layers.py
pointer only (licence: MIT); the stored copy get_code serves, from mshahbazi72/cGANTransfer, is pointer only (licence: NONE) · get_code("c41836f3e51fa8f1")
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 mshahbazi72/cGANTransfer, 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. A run is recorded.

loss_dcgan_dis Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/losses.py
code served (permissive licence) · get_code("982a26492dc6b7fb")
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. 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.functional as F

torch.manual_seed(0)

# Create small discriminator output tensors
# dis_fake: discriminator output for fake samples
dis_fake = torch.randn(2, 4)

# dis_real: discriminator output for real samples
dis_real = torch.randn(2, 4)

ARGS = {
    "dis_fake": dis_fake,
    "dis_real": dis_real,
}
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
loss_dcgan_gen Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/losses.py
code served (permissive licence) · get_code("6c34831223553b6f")
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 loss_dcgan_gen expects dis_fake, which is used with:
# - F.softplus() - expects a tensor
# - torch.mean() - expects a tensor
# So dis_fake should be a tensor

# Create a small fake discriminator output tensor (batch_size=2, some features)
dis_fake = torch.randn(2, 4)

ARGS = {
    "dis_fake": dis_fake,
}
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
loss_hinge_dis Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/losses.py
code served (permissive licence) · get_code("290cd37ab0098d5c")
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. 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.functional as F

torch.manual_seed(0)

# The function loss_hinge_dis takes two tensor arguments:
# dis_fake and dis_real
# It computes relu-based hinge losses on them and returns a tuple of two scalars

# Create small discriminator output tensors (batch_size=2, arbitrary values)
dis_fake = torch.randn(2, 4)  # Fake discriminator scores
dis_real = torch.randn(2, 4)  # Real discriminator scores

ARGS = {
    "dis_fake": dis_fake,
    "dis_real": dis_real,
}
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
pnorm Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/data_utils/resnet.py
code served (permissive licence) · get_code("2dda002928dada98")
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 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 torch

torch.manual_seed(0)

# Create a small tensor with shape (2, 4) - batch size 2, feature dimension 4
weights = torch.randn(2, 4)

# The function takes 'weights' and 'p' as parameters
# weights: a 2D tensor where it computes L2 norms along dimension 1
# p: a scalar that controls the power normalization

p = 2.0

ARGS = {
    "weights": weights,
    "p": p
}
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 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
power_iteration Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/layers.py
code served (permissive licence) · get_code("28d3a35157f2916d")
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
import torch.nn.functional as F

torch.manual_seed(0)

# Construct valid arguments for power_iteration
# W: weight matrix that will be multiplied with u and v vectors
# Shape should allow matmul operations: (4, 3) so u (1, 4) @ W (4, 3) = (1, 3)
W = torch.randn(4, 3)

# u_: list of initial vectors
# Each should be a 1D tensor with size matching W's first dimension
u_ = [
    torch.randn(1, 4),
    torch.randn(1, 4),
]

# update: boolean flag
update = True

# eps: small epsilon for numerical stability
eps = 1e-12

ARGS = {
    "W": W,
    "u_": u_,
    "update": update,
    "eps": eps,
}
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
resnet50 Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/data_utils/resnet.py
code served (permissive licence) · get_code("d404b2e50bcbb214")
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 ResNet_mine. 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
from torchvision.models.resnet import BasicBlock, Bottleneck, ResNet

torch.manual_seed(0)

ARGS = {
    'pretrained': False,
    'progress': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResNet_mine — a ResNet_mine output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet_mine 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
resnext50_32x4d Ran
ran; our contract draft was wrong, not the code
VectorInstitute/GenerativeSSL/data_generation/icgan/data_utils/resnet.py
code served (permissive licence) · get_code("a60b9f8ec1c25d5d")
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 ResNet_mine. 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
from torchvision.models.resnet import Bottleneck

torch.manual_seed(0)

ARGS = {
    'pretrained': False,
    'progress': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResNet_mine — a ResNet_mine output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet_mine 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
get_output Not yet run
unverified (harvested, no recorded run)
VectorInstitute/GenerativeSSL/data_generation/icgan/predict.py
code served (permissive licence) · get_code("567e5e16e8b1da99")
How it ran

no invocation recorded.

proj Not yet run
unverified (harvested, no recorded run)
VectorInstitute/GenerativeSSL/data_generation/icgan/BigGAN_PyTorch/layers.py
pointer only (licence: MIT); the stored copy get_code serves, from mshahbazi72/cGANTransfer, is pointer only (licence: NONE) · get_code("4015ab1f3ab9b881")
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 mshahbazi72/cGANTransfer, 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 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 11 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 0 the same month, 0 later) and 1 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
gram_schmidt 28 other papers
28 dated earlier · 5 only as an assembly brick
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1811.10597 (2018-11)
GAN Dissection: Visualizing and Understanding Generative Adversarial Networks
1910.05852 (2019-10)
Implicit competitive regularization in GANs
and 25 more not named here
proj 28 other papers
28 dated earlier · 5 only as an assembly brick
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1811.10597 (2018-11)
GAN Dissection: Visualizing and Understanding Generative Adversarial Networks
1910.05852 (2019-10)
Implicit competitive regularization in GANs
and 25 more not named here
power_iteration 25 other papers
25 dated earlier · 5 only as an assembly brick
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1811.10597 (2018-11)
GAN Dissection: Visualizing and Understanding Generative Adversarial Networks
1910.05852 (2019-10)
Implicit competitive regularization in GANs
and 22 more not named here
barlow_loss_func 12 other papers
12 dated earlier
2007.06346 (2020-07)
Whitening for Self-Supervised Representation Learning
2108.01775 (2021-08)
Solo-learn: A Library of Self-supervised Methods for Visual Representation Learning
2112.15022 (2021-12)
Continually Learning Self-Supervised Representations with Projected Functional Regularization
and 9 more not named here
loss_hinge_dis 11 other papers
11 dated earlier
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1912.00953 (2019-12)
LOGAN: Latent Optimisation for Generative Adversarial Networks
1912.05270 (2019-12)
MineGAN: effective knowledge transfer from GANs to target domains with few images
and 8 more not named here
loss_dcgan_dis 10 other papers
10 dated earlier
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
1912.05270 (2019-12)
MineGAN: effective knowledge transfer from GANs to target domains with few images
2008.10599 (2020-08)
The Hessian Penalty: A Weak Prior for Unsupervised Disentanglement
and 7 more not named here
loss_dcgan_gen 9 other papers
9 dated earlier
1809.11096 (2018-09)
Large Scale GAN Training for High Fidelity Natural Image Synthesis
2008.10599 (2020-08)
The Hessian Penalty: A Weak Prior for Unsupervised Disentanglement
2108.07668 (2021-08)
Orthogonal Jacobian Regularization for Unsupervised Disentanglement in Image Generation
and 6 more not named here
byol_loss_func 6 other papers
6 dated earlier
2007.06346 (2020-07)
Whitening for Self-Supervised Representation Learning
2204.00616 (2022-04)
Simplicial Embeddings in Self-Supervised Learning and Downstream Classification
2310.02903 (2023-10)
FroSSL: Frobenius Norm Minimization for Efficient Multiview Self-Supervised Learning
and 3 more not named here
pnorm 1 other paper
1 dated earlier
2412.06209 (2024-12)
Sound2Vision: Generating Diverse Visuals from Audio through Cross-Modal Latent Alignment
resnet50 1 other paper
1 dated earlier
2412.06209 (2024-12)
Sound2Vision: Generating Diverse Visuals from Audio through Cross-Modal Latent Alignment
resnext50_32x4d 1 other paper
1 dated earlier
2412.06209 (2024-12)
Sound2Vision: Generating Diverse Visuals from Audio through Cross-Modal Latent Alignment

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

The rapid advancement in self-supervised representation learning has highlighted its potential to leverage unlabeled data for learning rich visual representations. However, the existing techniques, particularly those employing different augmentations of the same image, often rely on a limited set of simple transformations that cannot fully capture variations in the real world. This constrains the diversity and quality of samples, which leads to sub-optimal representations. In this paper, we introduce a framework that enriches the self-supervised learning (SSL) paradigm by utilizing generative models to produce semantically consistent image augmentations. By directly conditioning generative models on a source image, our method enables the generation of diverse augmentations while maintaining the semantics of the source image, thus offering a richer set of data for SSL. Our extensive experimental results on various joint-embedding SSL techniques demonstrate that our framework significantly enhances the quality of learned visual representations by up to 10% Top-1 accuracy in downstream tasks. This research demonstrates that incorporating generative models into the joint-embedding SSL workflow opens new avenues for exploring the potential of synthetic data. This development paves the way for more robust and versatile representation learning techniques.

For agents

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

get_harvested_code_for_paper("aaai_32174")
get_code_for_paper("aaai_32174")
have("aaai_32174")

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

curl https://syntology.ai/api/ran/aaai_32174.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_32174.svg)](https://syntology.ai/paper/aaai_32174)

Connect an agent — have() is free.