SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.07366 · ICLR · 2024

Entropy is not Enough for Test-Time Adaptation: From the Perspective of Disentangled Factors

Sungroh Yoon, Saehyung Lee, Jonghyun Lee, Uiwon Hwang, Dahuin Jung, Junsung Park, Juhyeon Shin

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 3 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/2403.07366.json · badge: /api/ran/2403.07366.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
Jhyun17/DeYO canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
DeYO Ran
ran on a synthesized input · constructed an object; no result was computed
Jhyun17/DeYO/methods/deyo.py
code served (permissive licence) · get_code("e7c414643baf231f")
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 an instance of its own class, DeYO: the run constructed the object and did not call it, so no result was computed. 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 math

torch.manual_seed(0)

# Define a minimal model that satisfies the usage in forward_and_adapt_deyo
class MinimalModel(nn.Module):
    def __init__(self, num_classes=1000, input_channels=3):
        super().__init__()
        self.num_classes = num_classes
        self.fc = nn.Linear(input_channels * 4 * 4, num_classes)
    
    def forward(self, x):
        # x shape: (batch, channels, height, width)
        batch_size = x.shape[0]
        x = x.view(batch_size, -1)
        return self.fc(x)

# Create model and optimizer
model = MinimalModel(num_classes=1000, input_channels=3)
optimizer = torch.optim.SGD(model.parameters(), lr=0.001)

# Define args object with all required attributes
class Args:
    def __init__(self):
        self.filter_ent = True
        self.aug_type = 'occ'  # can be 'occ', 'patch', or 'pixel'
        self.occlusion_size = 2
        self.row_start = 0
        self.column_start = 0
        self.patch_len = 2
        self.filter_plpd = True
        self.plpd_threshold = 0.1
        self.reweight_ent = 0.5
        self.reweight_plpd = 0.5
        self.wandb_log = False

args = Args()

# Construct ARGS dictionary for DeYO.__init__
ARGS = {
    'model': model,
    'args': args,
    'optimizer': optimizer,
    'steps': 1,
    'episodic': False,
    'deyo_margin': 0.5 * math.log(1000),
    'margin_e0': 0.4 * math.log(1000)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch, torchvision, wandb — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DeYO — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DeYO 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
collect_params Ran
ran; our contract draft was wrong, not the code
Jhyun17/DeYO/methods/deyo.py
code served (permissive licence) · get_code("0c7e495c41c9465c")
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) 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 as nn

torch.manual_seed(0)

# Build a minimal model with normalization layers
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        # Add some norm layers that should be collected
        self.norm1 = nn.BatchNorm2d(4)
        self.norm2 = nn.LayerNorm(4)
        self.norm3 = nn.GroupNorm(2, 4)
        
        # Add some layers that should be skipped
        self.layer4 = nn.BatchNorm2d(4)
        self.blocks = nn.ModuleDict({
            '9': nn.BatchNorm2d(4),
            '10': nn.BatchNorm2d(4),
            '11': nn.BatchNorm2d(4),
        })
        
        # Add a norm. layer that should be skipped
        self.norm = nn.BatchNorm2d(4)
        
        # Add a regular layer
        self.linear = nn.Linear(4, 4)

model = SimpleModel()

ARGS = {
    'model': model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, 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
forward_and_adapt_deyo Ran
ran; our contract draft was wrong, not the code
Jhyun17/DeYO/methods/deyo.py
code served (permissive licence) · get_code("99ffb7c9f7a79773")
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 as nn
import math

torch.manual_seed(0)

# Create a minimal model that takes input and produces logits
class SimpleModel(nn.Module):
    def __init__(self, num_classes=1000):
        super().__init__()
        self.fc = nn.Linear(3 * 8 * 8, num_classes)
    
    def forward(self, x):
        # x shape: (batch, channels, height, width)
        batch_size = x.shape[0]
        x = x.view(batch_size, -1)
        return self.fc(x)

# Create a minimal args object with all required attributes
class Args:
    def __init__(self):
        self.filter_ent = True
        self.aug_type = 'occ'  # or 'patch', 'pixel'
        self.occlusion_size = 2
        self.row_start = 0
        self.column_start = 0
        self.patch_len = 2
        self.filter_plpd = True
        self.plpd_threshold = 0.1
        self.reweight_ent = 0.5
        self.reweight_plpd = 0.5

# Create test data
batch_size = 2
num_channels = 3
height = 8
width = 8
num_classes = 1000

x = torch.randn(batch_size, num_channels, height, width)
targets = torch.randint(0, num_classes, (batch_size,))
iter_ = 0
model = SimpleModel(num_classes=num_classes)
args = Args()
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
deyo_margin = math.log(500)
margin = math.log(100)

ARGS = {
    'x': x,
    'iter_': iter_,
    'model': model,
    'args': args,
    'optimizer': optimizer,
    'deyo_margin': deyo_margin,
    'margin': margin,
    'targets': targets,
    'flag': True,
    'group': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch, torchvision — 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
softmax_entropy Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from Feliciaxyao/ICML2024-FSTTA
pointer only; the stored copy get_code serves, from Feliciaxyao/ICML2024-FSTTA, is pointer only (licence: NONE) · get_code("3078648c75e3e1e3")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 3 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 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
softmax_entropy
brick of this paper's assembly: loss
22 other papers
11 dated earlier, 10 later, 1 undated · 4 only as an assembly brick
2006.10726 (2020-06)
TENT: FULLY TEST-TIME ADAPTATION BY ENTROPY MINIMIZATION
2105.08714 (2021-05)
Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks
2110.14038 (2021-10)
Robustness of Graph Neural Networks at Scale
and 19 more not named here
collect_params 1 other paper
1 undated
no dated paper among those named
openreview_QYPKLGyz7K
SaTeen: Learning Structural Alignment for Continual Test-Time Adaptation
forward_and_adapt_deyo
brick of this paper's assembly: update_rule
1 other paper
1 undated
no dated paper among those named
openreview_QYPKLGyz7K
SaTeen: Learning Structural Alignment for Continual Test-Time Adaptation

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

Test-time adaptation (TTA) fine-tunes pre-trained deep neural networks for unseen test data. The primary challenge of TTA is limited access to the entire test dataset during online updates, causing error accumulation. To mitigate it, TTA methods have utilized the model output's entropy as a confidence metric that aims to determine which samples have a lower likelihood of causing error. Through experimental studies, however, we observed the unreliability of entropy as a confidence metric for TTA under biased scenarios and theoretically revealed that it stems from the neglect of the influence of latent disentangled factors of data on predictions. Building upon these findings, we introduce a novel TTA method named Destroy Your Object (DeYO), which leverages a newly proposed confidence metric named Pseudo-Label Probability Difference (PLPD). PLPD quantifies the influence of the shape of an object on prediction by measuring the difference between predictions before and after applying an object-destructive transformation. DeYO consists of sample selection and sample weighting, which employ entropy and PLPD concurrently. For robust adaptation, DeYO prioritizes samples that dominantly incorporate shape information when making predictions. Our extensive experiments demonstrate the consistent superiority of DeYO over baseline methods across various scenarios, including biased and wild. Project page is publicly available at https://whitesnowdrop.github.io/DeYO/.

For agents

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

get_harvested_code_for_paper("2403.07366")
get_code_for_paper("2403.07366")
have("2403.07366")

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

curl https://syntology.ai/api/ran/2403.07366.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.07366.svg)](https://syntology.ai/paper/2403.07366)

Connect an agent — have() is free.