SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.02678 · ICML · 2024

Position: Quo Vadis, Unsupervised Time Series Anomaly Detection?

M Saquib Sarfraz, Mei-Yen Chen, Lukas Layer, Kunyu Peng, Marios Koulakis

arXiv · PDF · Open in the Atlas

Code that ran

14 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 13 failures of our instrument; 1 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/2405.02678.json · badge: /api/ran/2405.02678.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
ssarfraz/QuoVadisTAD canonical (classifier label)
classifier basis: census
14 ran · 1 no recorded run
FunctionStatusWhere it lives
add_anomalies_to_univariate_series Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/synthetic_series.py
code served (permissive licence) · get_code("ef109be6043f560e")
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 numpy as np
from scipy import stats

# Set random seed for reproducibility
np.random.seed(0)

# Create a simple univariate time series
x = np.sin(np.linspace(0, 4 * np.pi, 100)) + np.random.normal(0, 0.1, 100)

# Define parameters
normal_duration_rate = 5.0
anomaly_duration_rate = 3.0
anomaly_size_range = (0.5, 0.8)

ARGS = {
    "x": x,
    "normal_duration_rate": normal_duration_rate,
    "anomaly_duration_rate": anomaly_duration_rate,
    "anomaly_size_range": anomaly_size_range,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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
adj_standard Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/gnn.py
code served (permissive licence) · get_code("d49126dac29f6ae1")
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 2, dtype family float, shape (3, 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

# Create a small sample array for testing
# Shape: (features, samples) - will be transposed in the function to (samples, features)
# Using distinct dimensions: 4 features, 3 samples
train_array = np.random.randn(4, 3).astype(np.float32)

ARGS = {
    "train_array": train_array,
    "topk": 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 3) — shape (3, 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
compute_adj_on_slices Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/gnn.py
code served (permissive licence) · get_code("549caeddffb77285")
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 matrix, ndim 2, dtype family float, shape (16, 16) 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 scipy.sparse as sp

np.random.seed(0)

# Create a small test array with shape (num_samples, dim)
# num_samples should be > 2 * seq_len to allow slicing
# Using seq_len=16, we need at least 32 + 1 samples, so let's use 50
num_samples = 50
dim = 8
arr = np.random.randn(num_samples, dim).astype(np.float32)

ARGS = {
    "arr": arr,
    "seq_len": 16,
    "slices": 10,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type matrix, ndim 2, dtype family float, shape (16, 16) — shape (16, 16) 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
compute_finch_adjacency_matrix Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/gnn.py
code served (permissive licence) · get_code("7ba0eb523d055ae9")
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 lil_matrix. 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 numpy as np
from sklearn import metrics
import scipy.sparse as sp

# Create a small train_array with shape (timestamps, sensors)
# Function transposes it to (sensors, timestamps)
# Let's use timestamps=5, sensors=3
train_array = np.random.randn(5, 3).astype(np.float32)

ARGS = {
    "train_array": train_array,
    "symmetric": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type lil_matrix — a lil_matrix output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a lil_matrix 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
concatenate_windows_feat Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/data_utils.py
code served (permissive licence) · get_code("dfb4bb8f14b79db2")
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 2, dtype family float, shape (6, 20) 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

# Create a simple 2D numpy array with shape (batch_size, features)
# Using distinct sizes: 6 rows, 4 columns to make shapes unambiguous
arr = np.random.randn(6, 4).astype(np.float32)

# window_size parameter - keep it small but meaningful
window_size = 5

ARGS = {
    "arr": arr,
    "window_size": window_size,
}
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 2, dtype family float, shape (6, 20) — shape (6, 20) 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
create_tf_dataset Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/tf_data_loader.py
code served (permissive licence) · get_code("a971fa123110463b")
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 CacheDataset. 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 numpy as np
import tensorflow as tf

# Set random seed for reproducibility
np.random.seed(0)

# Create a small synthetic time series data array
# Shape: (num_time_steps, num_routes)
# Using 50 time steps and 4 routes
num_time_steps = 50
num_routes = 4
data_array = np.random.randn(num_time_steps, num_routes).astype(np.float32)

# Define parameters
input_sequence_length = 10
forecast_horizon = 5
batch_size = 2
reconstruction_dataset = False
shuffle = False
multi_horizon = False
input_for_gnn = False

ARGS = {
    "data_array": data_array,
    "input_sequence_length": input_sequence_length,
    "forecast_horizon": forecast_horizon,
    "batch_size": batch_size,
    "reconstruction_dataset": reconstruction_dataset,
    "shuffle": shuffle,
    "multi_horizon": multi_horizon,
    "input_for_gnn": input_for_gnn,
}
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 CacheDataset — a CacheDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a CacheDataset 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
find_files_in_path Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/dataset_reader.py
code served (permissive licence) · get_code("1bb7891844cde6e8")
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 list, 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 os
import tempfile
import shutil

# Create a temporary directory structure with some test files
temp_dir = tempfile.mkdtemp()

# Create some .npy files
open(os.path.join(temp_dir, 'file1.npy'), 'w').close()
open(os.path.join(temp_dir, 'file2.npy'), 'w').close()

# Create a subdirectory with more files
subdir = os.path.join(temp_dir, 'subdir')
os.makedirs(subdir, exist_ok=True)
open(os.path.join(subdir, 'file3.npy'), 'w').close()

# Create some non-.npy files that should be ignored
open(os.path.join(temp_dir, 'file4.txt'), 'w').close()
open(os.path.join(subdir, 'file5.csv'), 'w').close()

ARGS = {
    'directory': temp_dir,
    'file_ending': '.npy'
}
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 list, ndim 1, dtype family float — a list 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
get_label_loader Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/model_def.py
code served (permissive licence) · get_code("6f4a6cd9b2eb098a")
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 _BatchDataset. 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 numpy as np
from tensorflow.keras.preprocessing import timeseries_dataset_from_array

# Create small label data
# timeseries_dataset_from_array expects a 1D or 2D array
# We'll create a simple 1D array of shape (10,) to have some sequence data
labels = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0], dtype=np.float32)

# Create a config dict with required parameters
config = {
    'input_sequence_length': 3,
    'batch_size': 2
}

ARGS = {
    'labels': labels,
    'config': config
}
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 _BatchDataset — a _BatchDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a _BatchDataset 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
load_wadi_112 Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/dataset_reader.py
code served (permissive licence) · get_code("1caf05df591ca634")
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 numpy as np
from pathlib import Path
from unittest.mock import patch, MagicMock
import tempfile
import os

# Create a temporary directory structure that mimics the expected layout
temp_dir = tempfile.mkdtemp()
data_dir = Path(temp_dir) / 'resources' / 'processed_datasets' / 'WADI_112'
data_dir.mkdir(parents=True, exist_ok=True)

# Create small dummy numpy arrays and save them
train_data = np.random.randn(10, 5).astype(np.float32)
test_data = np.random.randn(8, 5).astype(np.float32)
test_labels = np.random.randint(0, 2, size=(8,)).astype(np.int32)

np.save(data_dir / 'train.npy', train_data)
np.save(data_dir / 'test.npy', test_data)
np.save(data_dir / 'labels.npy', test_labels)

# Construct the data_path argument
data_path = temp_dir

ARGS = {
    'data_path': data_path
}
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 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
load_wadi_127 Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/dataset_reader.py
code served (permissive licence) · get_code("68e07f6da439fc28")
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 numpy as np
import tempfile
from pathlib import Path

# Create temporary directory structure matching what the function expects
temp_dir = tempfile.mkdtemp()
temp_path = Path(temp_dir)
wadi_dir = temp_path / 'resources' / 'processed_datasets' / 'WADI_127'
wadi_dir.mkdir(parents=True, exist_ok=True)

# Create dummy numpy files with distinct shapes for train, test, and labels
train_data = np.random.rand(10, 5).astype(np.float32)  # 10 samples, 5 features
test_data = np.random.rand(4, 5).astype(np.float32)    # 4 samples, 5 features
test_labels = np.random.randint(0, 2, size=(4,)).astype(np.int32)  # 4 labels

# Save the files
np.save(wadi_dir / 'train.npy', train_data)
np.save(wadi_dir / 'test.npy', test_data)
np.save(wadi_dir / 'labels.npy', test_labels)

# Define ARGS with the data_path parameter
ARGS = {
    'data_path': temp_path
}
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 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
norm_1_1 Ran
ran, honoured the contract we drafted
ssarfraz/QuoVadisTAD/quovadis_tad/model_utils/model_def.py
code served (permissive licence) · get_code("7f7d95ada40f0907")
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
import numpy as np

torch.manual_seed(0)

# The function norm_1_1 takes an array and applies the transformation: 2 * arr - 1
# It expects a numeric array-like object (numpy array, torch tensor, or list)

# Create a small test array
arr = torch.tensor([[1.0, 2.0, 3.0, 4.0],
                     [5.0, 6.0, 7.0, 8.0]])

ARGS = {
    "arr": arr
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
normalise_scores Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/data_utils.py
code served (permissive licence) · get_code("774ce9c1ec908429")
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 2, dtype family float, shape (10, 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 numpy as np

# Create test_delta as a 2D numpy array with distinct dimensions
# Shape: (10, 4) - 10 time steps, 4 features
np.random.seed(0)
test_delta = np.random.randn(10, 4)

ARGS = {
    'test_delta': test_delta,
    'norm': 'median-iqr',
    'smooth': True,
    'smooth_window': 5,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (10, 4) — shape (10, 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
preprocess_data Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/data_utils.py
code served (permissive licence) · get_code("baeee7bf45f131bd")
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 numpy as np
from sklearn.preprocessing import StandardScaler, MinMaxScaler

# Create small sample data arrays
np.random.seed(0)
num_time_steps = 100
num_routes = 4

# Create data_array with shape (num_time_steps, num_routes)
data_array = np.random.randn(num_time_steps, num_routes).astype('float64')

# Create test_array with compatible shape
test_array = np.random.randn(20, num_routes).astype('float64')

# Define train_size and val_size as floats between 0.0 and 1.0
train_size = 0.7
val_size = 0.2

# Define normalization method
normalization = 'mean-std'

ARGS = {
    'data_array': data_array,
    'test_array': test_array,
    'train_size': train_size,
    'val_size': val_size,
    'normalization': normalization
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
synthetic_dataset_with_out_of_range_anomalies Ran
ran; our contract draft was wrong, not the code
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/synthetic_series.py
code served (permissive licence) · get_code("8abe71fff56ac4fc")
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 numpy as np
from scipy import stats

ARGS = {
    "number_of_sensors": 3,
    "train_size": 100,
    "test_size": 50,
    "nominal_data_mean": 0.0,
    "nominal_data_std": 0.1,
    "normal_duration_rate": 10.0,
    "anomaly_duration_rate": 5.0,
    "anomaly_size_range": (0.5, 0.8),
    "ratio_of_anomalous_sensors": 0.33,
    "seed": 42,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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
convert_array_to_dataframe_for_plotting Not yet run
unverified (harvested, no recorded run)
ssarfraz/QuoVadisTAD/quovadis_tad/dataset_utils/visualization.py
code served (permissive licence) · get_code("d1a6b21a4c3a9471")
How it ran

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, 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 15 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

The current state of machine learning scholarship in Timeseries Anomaly Detection (TAD) is plagued by the persistent use of flawed evaluation metrics, inconsistent benchmarking practices, and a lack of proper justification for the choices made in novel deep learning-based model designs. Our paper presents a critical analysis of the status quo in TAD, revealing the misleading track of current research and highlighting problematic methods, and evaluation practices. Our position advocates for a shift in focus from solely pursuing novel model designs to improving benchmarking practices, creating non-trivial datasets, and critically evaluating the utility of complex methods against simpler baselines. Our findings demonstrate the need for rigorous evaluation protocols, the creation of simple baselines, and the revelation that state-of-the-art deep anomaly detection models effectively learn linear mappings. These findings suggest the need for more exploration and development of simple and interpretable TAD methods. The increment of model complexity in the state-of-the-art deeplearning based models unfortunately offers very little improvement. We offer insights and suggestions for the field to move forward.

For agents

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

get_harvested_code_for_paper("2405.02678")
get_code_for_paper("2405.02678")
have("2405.02678")

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

curl https://syntology.ai/api/ran/2405.02678.json

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

[![Syntology run record](https://syntology.ai/api/ran/2405.02678.svg)](https://syntology.ai/paper/2405.02678)

Connect an agent — have() is free.