SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_25623 · AAAI · 2023

Detecting Multivariate Time Series Anomalies with Zero Known Label

Haoyu Liu, Qihang Zhou, Shibo He, Jiming Chen, Wenchao Meng

Open in the Atlas

Code that ran

7 of 16 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 1 has no contract verdict; 5 failures of our instrument; 9 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_25623.json · badge: /api/ran/aaai_25623.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
zqhang/MTGFLOW canonical (classifier label)
classifier basis: census
7 ran · 9 no recorded run
FunctionStatusWhere it lives
ConvDecoder Ran
ran; our contract draft was wrong, not the code
zqhang/MTGFLOW/models/GAN.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("91eab0777cc88db7")
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 wyy-code/MADGA, 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.

ROC Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zqhang/MTGFLOW/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("9ecd315dca81b217")
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 wyy-code/MADGA, 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.

R_Loss Ran
ran; our contract draft was wrong, not the code
zqhang/MTGFLOW/models/GAN.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("d2e490359ddee091")
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 wyy-code/MADGA, 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.

create_masks Ran
ran; our contract draft was wrong, not the code
zqhang/MTGFLOW/models/NF.py
pointer only (licence: MIT); the stored copy get_code serves, from enyandai/ganf, is pointer only (licence: NONE) · get_code("e1243b39051334f9")
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 enyandai/ganf, 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.

get_data_dim Ran
ran, honoured the contract we drafted
zqhang/MTGFLOW/Dataset/smd_smap_msl.py
code served (permissive licence) · get_code("f4c6bcd6ed01498a")
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 int, ndim 0, dtype family int, 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

ARGS = {
    'dataset': 'SMAP'
}
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 int, ndim 0, dtype family int, 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
interpolate Ran
ran; our contract draft was wrong, not the code
zqhang/MTGFLOW/models/MTGFLOW.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("7a35dac7185a5742")
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 wyy-code/MADGA, 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.

loader_SWat Ran
ran on a synthesized input
zqhang/MTGFLOW/Dataset/swat.py
code served (permissive licence) · get_code("2cbe939f52e0872a")
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 numpy as np
import pandas as pd
from io import StringIO

torch.manual_seed(0)

# Create a minimal synthetic SWat-like CSV data
csv_data = """Timestamp;Sensor_1;Sensor_2;Sensor_3;Sensor_4;Normal/Attack
2023-01-01 00:00:00;1.5;2.3;3.1;4.2;Normal
2023-01-01 00:00:01;1.6;2.4;3.2;4.3;Normal
2023-01-01 00:00:02;1.7;2.5;3.3;4.4;Normal
2023-01-01 00:00:03;1.8;2.6;3.4;4.5;Normal
2023-01-01 00:00:04;1.9;2.7;3.5;4.6;Normal
2023-01-01 00:00:05;2.0;2.8;3.6;4.7;Normal
2023-01-01 00:00:06;2.1;2.9;3.7;4.8;Normal
2023-01-01 00:00:07;2.2;3.0;3.8;4.9;Normal
2023-01-01 00:00:08;2.3;3.1;3.9;5.0;Normal
2023-01-01 00:00:09;2.4;3.2;4.0;5.1;Normal
2023-01-01 00:00:10;2.5;3.3;4.1;5.2;Normal
2023-01-01 00:00:11;2.6;3.4;4.2;5.3;Normal
2023-01-01 00:00:12;2.7;3.5;4.3;5.4;Normal
2023-01-01 00:00:13;2.8;3.6;4.4;5.5;Normal
2023-01-01 00:00:14;2.9;3.7;4.5;5.6;Normal
2023-01-01 00:00:15;3.0;3.8;4.6;5.7;Normal
2023-01-01 00:00:16;3.1;3.9;4.7;5.8;Normal
2023-01-01 00:00:17;3.2;4.0;4.8;5.9;Normal
2023-01-01 00:00:18;3.3;4.1;4.9;6.0;Normal
2023-01-01 00:00:19;3.4;4.2;5.0;6.1;Normal
2023-01-01 00:00:20;3.5;4.3;5.1;6.2;Normal
2023-01-01 00:00:21;3.6;4.4;5.2;6.3;Normal
2023-01-01 00:00:22;3.7;4.5;5.3;6.4;Attack
2023-01-01 00:00:23;3.8;4.6;5.4;6.5;Attack
2023-01-01 00:00:24;3.9;4.7;5.5;6.6;Attack
2023-01-01 00:00:25;4.0;4.8;5.6;6.7;Normal
2023-01-01 00:00:26;4.1;4.9;5.7;6.8;Normal
2023-01-01 00:00:27;4.2;5.0;5.8;6.9;Normal
2023-01-01 00:00:28;4.3;5.1;5.9;7.0;Normal
2023-01-01 00:00:29;4.4;5.2;6.0;7.1;Normal
2023-01-01 00:00:30;4.5;5.3;6.1;7.2;Normal
"""

# Write to a temporary file-like object and create a real file
import tempfile
import os

temp_dir = tempfile.mkdtemp()
csv_file = os.path.join(temp_dir, "swat_data.csv")

with open(csv_file, 'w') as f:
    f.write(csv_data)

ARGS = {
    'root': csv_file,
    'batch_size': 2,
    'window_size': 5,
    'stride_size': 1,
    'train_split': 0.7,
    'label': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, sklearn, 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
ConvEncoder Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/models/GAN.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("0a503f99c3cb7063")
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 wyy-code/MADGA, 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.

create_masks_pmu Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/models/NF.py
pointer only (licence: MIT); the stored copy get_code serves, from wyy-code/MADGA, is pointer only (licence: NONE) · get_code("7f9163461fb01c53")
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 wyy-code/MADGA, 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.

load_smd_smap_msl Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/smd_smap_msl.py
code served (permissive licence) · get_code("03f7dc1a8b481933")
How it ran

no invocation recorded.

load_smd_smap_msl_occ Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/smd_smap_msl.py
code served (permissive licence) · get_code("5bf6fd1726d343ff")
How it ran

no invocation recorded.

loader_PSM Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/psm.py
code served (permissive licence) · get_code("4c48c398df8516b0")
How it ran

no invocation recorded.

loader_PSM_OCC Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/psm.py
code served (permissive licence) · get_code("aa10af60eb9a448c")
How it ran

no invocation recorded.

loader_SWat_OCC Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/swat.py
code served (permissive licence) · get_code("9a4516ca40cb1c9e")
How it ran

no invocation recorded.

loader_WADI Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/wadi.py
code served (permissive licence) · get_code("ee55bd9bfb9d3916")
How it ran

no invocation recorded.

loader_WADI_OCC Not yet run
unverified (harvested, no recorded run)
zqhang/MTGFLOW/Dataset/wadi.py
code served (permissive licence) · get_code("05affd7211c8c8d2")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 8 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 8 later) and 8 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
get_data_dim 3 other papers
2 dated earlier, 1 later
2009.02040 (2020-09)
Multivariate Time-series Anomaly Detection via Graph Attention Network
2108.06816 (2021-08)
Weakly Supervised Temporal Anomaly Segmentation with Dynamic Time Warping
2605.14632 (2026-05)
DRL-STAF: A Deep Reinforcement Learning Framework for State-Aware Forecasting of Complex Multivariate Hidden Markov Processes
create_masks 2 other papers
1 dated earlier, 1 later
2202.07857 (2022-02)
Graph-Augmented Normalizing Flows for Anomaly Detection of Multiple Time Series
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
ConvDecoder 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
ConvEncoder 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
ROC 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
R_Loss 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
create_masks_pmu 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
interpolate 1 other paper
1 later
2410.08877 (2024-10)
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection

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

Multivariate time series anomaly detection has been extensively studied under the one-class classification setting, where a training dataset with all normal instances is required. However, preparing such a dataset is very laborious since each single data instance should be fully guaranteed to be normal. It is, therefore, desired to explore multivariate time series anomaly detection methods based on the dataset without any label knowledge. In this paper, we propose MTGFlow, an unsupervised anomaly detection approach for Multivariate Time series anomaly detection via dynamic Graph and entityaware normalizing Flow, leaning only on a widely accepted hypothesis that abnormal instances exhibit sparse densities than the normal. However, the complex interdependencies among entities and the diverse inherent characteristics of each entity pose significant challenges to density estimation, let alone to detect anomalies based on the estimated possibility distribution. To tackle these problems, we propose to learn the mutual and dynamic relations among entities via a graph structure learning model, which helps to model the accurate distribution of multivariate time series. Moreover, taking account of distinct characteristics of the individual entities, an entity-aware normalizing flow is developed to describe each entity into a parameterized normal distribution, thereby producing fine-grained density estimation. Incorporating these two strategies, MTGFlow achieves superior anomaly detection performance. Experiments on five public datasets with seven baselines are conducted, MTGFlow outperforms the SOTA methods by up to 5.0 AUROC%.

For agents

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

get_harvested_code_for_paper("aaai_25623")
get_code_for_paper("aaai_25623")
have("aaai_25623")

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

curl https://syntology.ai/api/ran/aaai_25623.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_25623.svg)](https://syntology.ai/paper/aaai_25623)

Connect an agent — have() is free.