SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.06119 · 2023

Exploring Progress in Multivariate Time Series Forecasting: Comprehensive Benchmarking and Heterogeneity Analysis

arXiv · PDF · Open in the Atlas

Code that ran

11 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; 11 failures of our instrument; 1 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/2310.06119.json · badge: /api/ran/2310.06119.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
zezhishao/basicts canonical (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
zezhishao/d2stgnn canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
hitplz/dstrformer community (classifier label)
classifier basis not recorded on the link
2 ran · 1 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran; our contract draft was wrong, not the code
zezhishao/basicts/src/basicts/metrics/cls_metrics.py
code served (permissive licence) · get_code("b950e7ac7e9954a3")
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

torch.manual_seed(0)

# Create small tensor predictions (batch_size=2, num_classes=4)
# These will be class indices or binary predictions
prediction = torch.tensor([[0, 1, 2, 3],
                           [1, 0, 3, 2]], dtype=torch.long)

# Create target tensor with same shape
targets = torch.tensor([[0, 1, 2, 3],
                        [1, 0, 3, 2]], dtype=torch.long)

ARGS = {
    "prediction": prediction,
    "targets": targets,
}
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
l1_loss Ran
ran; our contract draft was wrong, not the code
hitplz/dstrformer/basicts/losses/losses.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from jimmy-7664/std-mae, is pointer only (licence: NONE) · get_code("48469f10c13c505c")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from jimmy-7664/std-mae, 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.

masked_corr Ran
ran; our contract draft was wrong, not the code
zezhishao/basicts/src/basicts/metrics/corr.py
code served (permissive licence) · get_code("f916c347df4200df")
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

torch.manual_seed(0)

# Create small tensors for testing
# batch_size=2, sequence_length=4
batch_size = 2
seq_length = 4

prediction = torch.randn(batch_size, seq_length)
targets = torch.randn(batch_size, seq_length)
targets_mask = torch.ones(batch_size, seq_length)

ARGS = {
    "prediction": prediction,
    "targets": targets,
    "targets_mask": targets_mask,
}
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
masked_huber Ran
ran; our contract draft was wrong, not the code
zezhishao/basicts/src/basicts/metrics/huber.py
code served (permissive licence) · get_code("968bb6c4da63389d")
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
from torch.nn import HuberLoss

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
feature_dim = 4

prediction = torch.randn(batch_size, feature_dim)
targets = torch.randn(batch_size, feature_dim)
targets_mask = torch.ones(batch_size, feature_dim)

ARGS = {
    'prediction': prediction,
    'targets': targets,
    'targets_mask': targets_mask,
    'reduction': 'mean',
    'delta': 1.0
}
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
masked_mae Ran
ran; our contract draft was wrong, not the code
zezhishao/basicts/src/basicts/metrics/mae.py
code served (permissive licence) · get_code("930f004619ca0a81")
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

torch.manual_seed(0)

# Create small tensors for prediction and targets
batch_size = 2
feature_dim = 4

prediction = torch.randn(batch_size, feature_dim)
targets = torch.randn(batch_size, feature_dim)
targets_mask = torch.ones(batch_size, feature_dim)

# Some entries can be masked out (set to 0)
targets_mask[0, 1] = 0.0

ARGS = {
    "prediction": prediction,
    "targets": targets,
    "targets_mask": targets_mask,
}
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
masked_mae Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
hitplz/dstrformer/basicts/losses/losses.py
code served (permissive licence) · get_code("0706b9df9b6a5e94")
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 numpy as np

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
dim = 4

# Create predictions tensor
preds = torch.randn(batch_size, dim)

# Create labels tensor with some NaN values to test masking
labels = torch.randn(batch_size, dim)
labels[0, 1] = np.nan  # Add a NaN value to test NaN masking

# null_val parameter - using np.nan as default
null_val = np.nan

ARGS = {
    "preds": preds,
    "labels": labels,
    "null_val": null_val,
}
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 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
masked_mae_loss Ran
ran; our contract draft was wrong, not the code
zezhishao/d2stgnn/models/losses.py
pointer only (licence: NONE); the stored copy get_code serves, from lingbai-kong/causalformer, is pointer only (licence: GPL-3.0) · get_code("2a82a90d91ced4b6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zezhishao/d2stgnn 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.

masked_mape Ran
ran; our contract draft was wrong, not the code
zezhishao/basicts/src/basicts/metrics/mape.py
code served (permissive licence) · get_code("636d3e348fb2bcd3")
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

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
dim = 4

# prediction: predicted values tensor
prediction = torch.randn(batch_size, dim, dtype=torch.float32)

# targets: ground truth values tensor with same shape as prediction
targets = torch.randn(batch_size, dim, dtype=torch.float32)
# Ensure targets are not all close to zero to avoid division issues
targets = targets + 2.0

# targets_mask: optional mask tensor with same shape as targets
# Using None to test the default behavior, or could be a binary mask
targets_mask = torch.ones(batch_size, dim, dtype=torch.float32)

ARGS = {
    "prediction": prediction,
    "targets": targets,
    "targets_mask": targets_mask,
}
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
masked_mape_np Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zezhishao/d2stgnn/models/trainer.py
pointer only (licence: NONE) · get_code("90560ad3b6c50922")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zezhishao/d2stgnn 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.

masked_mse Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zezhishao/d2stgnn/models/losses.py
pointer only (licence: NONE) · get_code("38743922b798588e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zezhishao/d2stgnn 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.

masked_rmse Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zezhishao/d2stgnn/models/losses.py
pointer only (licence: NONE) · get_code("98046566c0d9872a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zezhishao/d2stgnn 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.

l2_loss Not yet run
unverified (harvested, no recorded run)
hitplz/dstrformer/basicts/losses/losses.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from jimmy-7664/std-mae, is pointer only (licence: NONE) · get_code("e97291c1cbf62cfd")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from jimmy-7664/std-mae, 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 (1 with at least one paper dated earlier, 0 the same month, 11 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
masked_mae_loss 5 other papers
1 dated earlier, 3 later, 1 same year, order not determinable
aaai_25976 (2023 (year only))
Spatio-Temporal Meta-Graph Learning for Traffic Forecasting
2304.14343 (2023-04)
LibCity: A Unified Library Towards Efficient and Comprehensive Urban Spatial-Temporal Prediction
2312.11198 (2023-12)
Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-time Dynamics
and 2 more not named here
masked_rmse 2 other papers
2 later
2410.10524 (2024-10)
Get Rid of Isolation: A Continuous Multi-task Spatio-Temporal Learning Framework
2410.19192 (2024-10)
TEAM: Topological Evolution-aware Framework for Traffic Forecasting--Extended Version
accuracy 1 other paper
1 later
2601.12083 (2026-01)
Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models
l1_loss 1 other paper
1 later
2312.00516 (2023-12)
Spatial-Temporal-Decoupled Masked Pre-training for Spatiotemporal Forecasting
l2_loss 1 other paper
1 later
2312.00516 (2023-12)
Spatial-Temporal-Decoupled Masked Pre-training for Spatiotemporal Forecasting
masked_corr 1 other paper
1 later
2601.12083 (2026-01)
Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models
masked_huber 1 other paper
1 later
2601.12083 (2026-01)
Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models
masked_mae 1 other paper
1 later
2601.12083 (2026-01)
Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models
masked_mae 1 other paper
1 later
2405.11333 (2024-05)
GinAR: An End-To-End Multivariate Time Series Forecasting Model Suitable for Variable Missing
masked_mape 1 other paper
1 later
2601.12083 (2026-01)
Learning to Factorize and Adapt: A Versatile Approach Toward Universal Spatio-Temporal Foundation Models
masked_mse 1 other paper
1 later
2410.10524 (2024-10)
Get Rid of Isolation: A Continuous Multi-task Spatio-Temporal Learning Framework

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 (MTS) analysis is crucial to understanding and managing complex systems, such as traffic and energy systems, and a variety of approaches to MTS forecasting have been proposed recently. However, we often observe inconsistent or seemingly contradictory performance findings across different studies. This hinders our understanding of the merits of different approaches and slows down progress. We address the need for means of assessing MTS forecasting proposals reliably and fairly, in turn enabling better exploitation of MTS as seen in different applications. Specifically, we first propose BasicTS+, a benchmark designed to enable fair, comprehensive, and reproducible comparison of MTS forecasting solutions. BasicTS+ establishes a unified training pipeline and reasonable settings, enabling an unbiased evaluation. Second, we identify the heterogeneity across different MTS as an important consideration and enable classification of MTS based on their temporal and spatial characteristics. Disregarding this heterogeneity is a prime reason for difficulties in selecting the most promising technical directions. Third, we apply BasicTS+ along with rich datasets to assess the capabilities of more than 45 MTS forecasting solutions. This provides readers with an overall picture of the cutting-edge research on MTS forecasting. The code can be accessed at https://github.com/GestaltCogTeam/BasicTS.

For agents

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

get_harvested_code_for_paper("2310.06119")
get_code_for_paper("2310.06119")
have("2310.06119")

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

curl https://syntology.ai/api/ran/2310.06119.json

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

[![Syntology run record](https://syntology.ai/api/ran/2310.06119.svg)](https://syntology.ai/paper/2310.06119)

Connect an agent — have() is free.