SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2005.03788 · 2020

ProSelfLC: Progressive Self Label Correction for Training Robust Deep Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

9 of 11 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; 1 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 3 failures of our instrument; 2 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/2005.03788.json · badge: /api/ran/2005.03788.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
kiyoon/pyvideoai community (classifier label)
classifier basis not recorded on the link
9 ran · 2 no recorded run
FunctionStatusWhere it lives
count_class_frequency Ran
ran on a synthesized input
kiyoon/pyvideoai/pyvideoai/dataloaders/frames_sparsesample_dataset.py
code served (permissive licence) · get_code("4ccacd932d6aa3b7")
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 ndarray, ndim 1, dtype family int, shape (3,) 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 numpy as np
import os
import tempfile

# Create a temporary CSV file with valid format
temp_dir = tempfile.mkdtemp()
csv_file = os.path.join(temp_dir, 'test_labels.csv')

# Write a valid CSV file
with open(csv_file, 'w') as f:
    # First line: number of classes
    f.write('3\n')
    # Following lines: path video_id label start_frame end_frame
    f.write('video1.mp4 vid_1 0 10 20\n')
    f.write('video2.mp4 vid_2 1 15 25\n')
    f.write('video3.mp4 vid_3 2 5 15\n')
    f.write('video4.mp4 vid_4 0 30 40\n')
    f.write('video5.mp4 vid_5 1 20 30\n')

ARGS = {
    'csv_file': csv_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (3,) — shape (3,) 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
get_lr Ran
ran on a synthesized input
kiyoon/pyvideoai/pyvideoai/train_and_eval.py
code served (permissive licence) · get_code("532347378db99392")
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
from torch.optim import SGD

torch.manual_seed(0)

# Create a minimal model
model = nn.Linear(4, 3)

# Create an optimizer with one parameter group
optimiser = SGD(model.parameters(), lr=0.001)

ARGS = {
    'optimiser': optimiser
}
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
get_scheduler_type Ran
ran on a synthesized input
kiyoon/pyvideoai/pyvideoai/train_and_eval.py
code served (permissive licence) · get_code("c366aa3be5213eaa")
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 str. 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
from torch.optim.lr_scheduler import ReduceLROnPlateau

torch.manual_seed(0)

# Define the missing classes that the function references
class ReduceLROnPlateauMultiple:
    """Stand-in for ReduceLROnPlateauMultiple scheduler"""
    pass

class GradualWarmupScheduler:
    """Stand-in for GradualWarmupScheduler"""
    def __init__(self, after_scheduler=None):
        self.after_scheduler = after_scheduler

# Create a simple optimizer for the scheduler
model = nn.Linear(4, 2)
optimizer = torch.optim.SGD(model.parameters(), lr=0.1)

# Test case 1: None scheduler
scheduler_none = None

# Test case 2: ReduceLROnPlateau scheduler
scheduler_plateau = ReduceLROnPlateau(optimizer, mode='min', factor=0.1, patience=10)

# Test case 3: ReduceLROnPlateauMultiple scheduler
scheduler_plateau_multiple = ReduceLROnPlateauMultiple()

# Test case 4: GradualWarmupScheduler with ReduceLROnPlateau after_scheduler
warmup_with_plateau = GradualWarmupScheduler(after_scheduler=scheduler_plateau)

# Test case 5: GradualWarmupScheduler with ReduceLROnPlateauMultiple after_scheduler
warmup_with_plateau_multiple = GradualWarmupScheduler(after_scheduler=scheduler_plateau_multiple)

# Test case 6: GradualWarmupScheduler with no after_scheduler (or other scheduler)
warmup_with_none = GradualWarmupScheduler(after_scheduler=None)

# Use the first test case for ARGS
ARGS = {
    'scheduler': scheduler_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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_start_end_idx Ran
ran, honoured the contract we drafted
kiyoon/pyvideoai/pyvideoai/dataloaders/decoder.py
code served (permissive licence) · get_code("c1a3191abe820a2c")
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 random

# Set seed for reproducibility
random.seed(0)

# Based on the function signature and body, we need:
# - video_size: int - total number of frames in video
# - clip_size: int - number of frames to extract
# - clip_idx: int - index of clip (-1 for random, else 0 to num_clips-1)
# - num_clips: int - total number of clips to divide video into

ARGS = {
    "video_size": 100,
    "clip_size": 16,
    "clip_idx": 2,
    "num_clips": 10,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
register_vcs_handler Ran
ran; our contract draft was wrong, not the code
kiyoon/pyvideoai/pyvideoai/_version.py
code served (permissive licence) · get_code("b06b3b57c757198b")
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 function. 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

HANDLERS = {}

def register_vcs_handler(vcs, method):
    """Create decorator to mark a method as the handler of a VCS."""

    def decorate(f):
        """Store f in HANDLERS[vcs][method]."""
        if vcs not in HANDLERS:
            HANDLERS[vcs] = {}
        HANDLERS[vcs][method] = f
        return f
    return decorate


def sample_handler_function():
    """A simple handler function to decorate."""
    pass


ARGS = {
    "vcs": "git",
    "method": "clone"
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
run_command Ran
ran on a synthesized input
kiyoon/pyvideoai/pyvideoai/_version.py
code served (permissive licence) · get_code("c568a8377abe39e7")
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 subprocess
import sys

# The function runs a subprocess command. We need to provide:
# - commands: list of command strings to try
# - args: list of arguments to pass to the command
# - cwd: working directory (optional)
# - verbose: whether to print debug info (optional)
# - hide_stderr: whether to hide stderr (optional)
# - env: environment variables dict (optional)

# We'll use a simple, universally available command like 'echo'
ARGS = {
    'commands': ['echo'],
    'args': ['hello', 'world'],
    'cwd': None,
    'verbose': False,
    'hide_stderr': False,
    'env': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
temporal_sampling Ran
ran; our contract draft was wrong, not the code
kiyoon/pyvideoai/pyvideoai/dataloaders/decoder.py
code served (permissive licence) · get_code("2ce6782a49b98214")
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 4, dtype family float, shape (5, 3, 4, 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 tensor of video frames: num_frames x channels x height x width
frames = torch.randn(10, 3, 4, 4)

start_idx = 0
end_idx = 9
num_samples = 5

ARGS = {
    "frames": frames,
    "start_idx": start_idx,
    "end_idx": end_idx,
    "num_samples": num_samples,
}
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 4, dtype family float, shape (5, 3, 4, 4) — shape (5, 3, 4, 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
turn_absolute Ran
ran on a synthesized input
kiyoon/pyvideoai/pyvideoai/config.py
code served (permissive licence) · get_code("4adeea8f6880129b")
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 str. 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 os
import tempfile

# Create a temporary directory to act as the script directory
_temp_dir = tempfile.mkdtemp()

# Create a test file to serve as __file__
_test_file = os.path.join(_temp_dir, "test_script.py")
with open(_test_file, "w") as f:
    f.write("# test script")

# Simulate the module-level _SCRIPT_DIR that would be set in the actual module
_SCRIPT_DIR = os.path.dirname(os.path.realpath(_test_file))

def turn_absolute(path):
    if not os.path.isabs(path):
        return os.path.realpath(os.path.join(_SCRIPT_DIR, path))
    else:
        return path

# Test cases for the function
relative_path = "subdir/file.txt"
absolute_path = os.path.abspath("/tmp/absolute_file.txt")

ARGS = {
    "path": relative_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
versions_from_parentdir Ran
ran; our contract draft was wrong, not the code
kiyoon/pyvideoai/pyvideoai/_version.py
pointer only (licence: MIT); the stored copy get_code serves, from Hhuangsj/GREX, is pointer only (licence: NONE) · get_code("b9a1535d27f2a9fb")
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 Hhuangsj/GREX, 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.

pyav_decode_stream Not yet run
unverified (harvested, no recorded run)
kiyoon/pyvideoai/pyvideoai/dataloaders/decoder.py
code served (permissive licence) · get_code("b28b46ef7776c434")
How it ran

no invocation recorded.

train_iter Not yet run
unverified (harvested, no recorded run)
kiyoon/pyvideoai/pyvideoai/train_and_eval.py
code served (permissive licence) · get_code("db3b71c5b527f005")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 6 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 3 the same month, 5 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
register_vcs_handler 26 other papers
7 dated earlier, 1 the same month, 18 later
1806.00550 (2018-06)
A Swiss Army Infinitesimal Jackknife
1806.03146 (2018-06)
Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials
1806.03720 (2018-06)
Stochastic seismic waveform inversion using generative adversarial networks as a geological prior
and 23 more not named here
versions_from_parentdir 16 other papers
6 dated earlier, 1 the same month, 9 later
1806.00550 (2018-06)
A Swiss Army Infinitesimal Jackknife
1806.03720 (2018-06)
Stochastic seismic waveform inversion using generative adversarial networks as a geological prior
1812.01719 (2018-12)
Knowing what you know in brain segmentation using Bayesian deep neural networks
and 13 more not named here
pyav_decode_stream 4 other papers
4 later
Shi_Multi-Modal_Multi-Action_Video_Recognition_ICCV_2021_paper (2021 (year only))
Multi-Modal Multi-Action Video Recognition
136950157 (2022 (year only))
K-centered Patch Sampling for Efficient Video Recognition
2206.03428 (2022-06)
Revealing Single Frame Bias for Video-and-Language Learning
and 1 more not named here
temporal_sampling 4 other papers
4 later
Shi_Multi-Modal_Multi-Action_Video_Recognition_ICCV_2021_paper (2021 (year only))
Multi-Modal Multi-Action Video Recognition
136950157 (2022 (year only))
K-centered Patch Sampling for Efficient Video Recognition
2306.07703 (2023-06)
E2E-LOAD: End-to-End Long-form Online Action Detection
and 1 more not named here
get_start_end_idx 3 other papers
3 later
136640364 (2022 (year only))
Continual 3D Convolutional Neural Networks for Real-time Processing of Videos
2206.03428 (2022-06)
Revealing Single Frame Bias for Video-and-Language Learning
2303.04995 (2023-03)
Text-Visual Prompting for Efficient 2D Temporal Video Grounding
run_command 1 other paper
1 the same month
2005.08131 (2020-05)
Evaluation of synthetic and experimental training data in supervised machine learning applied to charge state detection of quantum dots

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

To train robust deep neural networks (DNNs), we systematically study several target modification approaches, which include output regularisation, self and non-self label correction (LC). Two key issues are discovered: (1) Self LC is the most appealing as it exploits its own knowledge and requires no extra models. However, how to automatically decide the trust degree of a learner as training goes is not well answered in the literature? (2) Some methods penalise while the others reward low-entropy predictions, prompting us to ask which one is better? To resolve the first issue, taking two well-accepted propositions--deep neural networks learn meaningful patterns before fitting noise [3] and minimum entropy regularisation principle [10]--we propose a novel end-to-end method named ProSelfLC, which is designed according to learning time and entropy. Specifically, given a data point, we progressively increase trust in its predicted label distribution versus its annotated one if a model has been trained for enough time and the prediction is of low entropy (high confidence). For the second issue, according to ProSelfLC, we empirically prove that it is better to redefine a meaningful low-entropy status and optimise the learner toward it. This serves as a defence of entropy minimisation. We demonstrate the effectiveness of ProSelfLC through extensive experiments in both clean and noisy settings. The source code is available at https://github.com/XinshaoAmosWang/ProSelfLC-CVPR2021. Keywords: entropy minimisation, maximum entropy, confidence penalty, self knowledge distillation, label correction, label noise, semi-supervised learning, output regularisation

For agents

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

get_harvested_code_for_paper("2005.03788")
get_code_for_paper("2005.03788")
have("2005.03788")

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

curl https://syntology.ai/api/ran/2005.03788.json

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

[![Syntology run record](https://syntology.ai/api/ran/2005.03788.svg)](https://syntology.ai/paper/2005.03788)

Connect an agent — have() is free.