SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.14286 · NeurIPS · 2023

Equivariant Neural Simulators for Stochastic Spatiotemporal Dynamics

Vlado Menkovski, Koen Minartz, Yoeri Poels, Simon Koop

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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 5 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2305.14286.json · badge: /api/ran/2305.14286.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
kminartz/EPNS canonical (classifier label)
classifier basis: census
5 ran · 0 no recorded run
FunctionStatusWhere it lives
arr_to_im Ran
ran on a synthesized input
kminartz/EPNS/utils.py
code served (permissive licence) · get_code("87cc1a538ca9f863")
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 3, dtype family float, shape (4, 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

# Test case 1: one-hot encoded input
# Shape should be (num_classes, height, width)
# The function checks arr.shape[-3], arr.shape[-2:], and uses arr[i] == 1.0
arr_one_hot = np.zeros((5, 4, 3))  # 5 classes, 4x3 spatial dims
# Set some pixels to 1.0 for each class
arr_one_hot[0, 0, 0] = 1.0
arr_one_hot[1, 1, 1] = 1.0
arr_one_hot[2, 2, 2] = 1.0

# Test case 2: non-one-hot input
# Shape should be (3, height, width) for RGB or similar
# The function does moveaxis(arr, (0, 1, 2), (2, 0, 1))
arr_regular = np.random.rand(3, 4, 5)  # 3 channels, 4x5 spatial dims

ARGS = {
    "arr": arr_one_hot,
    "is_one_hot_enc": True
}
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 3, dtype family float, shape (4, 3, 3) — shape (4, 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
get_data_loaders Ran
ran on a synthesized input
kminartz/EPNS/datasets/Cell_Combined_Dataset.py
code served (permissive licence) · get_code("87bab19f49c868f7")
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 os
import numpy as np
import torch
import tempfile
import shutil

torch.manual_seed(0)

# Create temporary directory structure with mock data
temp_dir = tempfile.mkdtemp()

# Create subdirectories
train_dir = os.path.join(temp_dir, 'train')
val_dir = os.path.join(temp_dir, 'val')
test_dir = os.path.join(temp_dir, 'test')
test_same_init_dir = os.path.join(temp_dir, 'test_same_init')

os.makedirs(train_dir)
os.makedirs(val_dir)
os.makedirs(test_dir)
os.makedirs(test_same_init_dir)

# Create mock numpy files with shape (timesteps=5, channels=3, height=4, width=4)
for i in range(2):
    np.save(os.path.join(train_dir, f'train_{i}.npy'), np.random.randn(5, 3, 4, 4))
    np.save(os.path.join(val_dir, f'val_{i}.npy'), np.random.randn(5, 3, 4, 4))
    np.save(os.path.join(test_dir, f'test_{i}.npy'), np.random.randn(5, 3, 4, 4))
    np.save(os.path.join(test_same_init_dir, f'test_same_{i}.npy'), np.random.randn(5, 3, 4, 4))

config = {
    'data_directory': temp_dir,
    'batch_size': 2
}

ARGS = {
    'config': config,
    'additional_loaders': ('test_same_init',),
    'limit_num_data_points_to': np.inf
}
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 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
get_data_loaders Ran
ran on a synthesized input
kminartz/EPNS/datasets/N_Body_Dataset.py
code served (permissive licence) · get_code("a77c39a49bdc63f5")
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 os
import tempfile
import warnings
import numpy as np
import torch
from torch.utils.data import Dataset

torch.manual_seed(0)

# Create a temporary directory structure with mock data
temp_dir = tempfile.mkdtemp()
for split in ['train', 'val', 'test', 'test_same_init', 'test_long_simulation']:
    split_dir = os.path.join(temp_dir, split)
    os.makedirs(split_dir, exist_ok=True)
    
    # Create 3 mock .npy files per split
    for i in range(3):
        # Shape: (num_bodies, features, timesteps)
        data = np.random.randn(5, 4, 10).astype(np.float32)
        np.save(os.path.join(split_dir, f'sample_{i}.npy'), data)

# Create config dict
config = {
    'data_directory': temp_dir,
    'batch_size': 2
}

ARGS = {
    'config': config,
    'additional_loaders': ('test_same_init', 'test_long_simulation'),
    'limit_num_data_points_to': np.inf,
    'expand_train_set': True
}
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 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
plot_NS_image Ran
ran on a synthesized input
kminartz/EPNS/utils.py
code served (permissive licence) · get_code("ee1f8489cb2b747b")
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 3, dtype family float, shape (4, 13, 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 matplotlib.pyplot as plt

# Create a simple datapoint with batch dimension
# The function expects datapoint[0] to work, so batch size should be >= 1
# arr_to_im expects either:
# - One-hot encoded: shape (num_channels, height, width) where channels are binary
# - Regular image: shape (channels, height, width) with values 0-1
# Based on the code, when not one-hot, it does moveaxis(arr, (0,1,2), (2,0,1))
# This converts (C, H, W) to (H, W, C)

# Create a batch of 2 datapoints, each with shape (3, 4, 5) for (channels, height, width)
datapoint = np.random.rand(2, 3, 4, 5)

# Create predicted output with same shape as datapoint
predicted = np.random.rand(2, 3, 4, 5)

# Define all arguments
ARGS = {
    'datapoint': datapoint,
    'predicted': predicted,
    'title_append': 'test_comparison',
    'no_title': False,
    'show_plot': False  # Don't actually display the plot during testing
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (4, 13, 3) — shape (4, 13, 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
plot_cell_image Ran
ran on a synthesized input
kminartz/EPNS/utils.py
code served (permissive licence) · get_code("4a00a6ed6a166e25")
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 3, dtype family float, shape (4, 13, 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 matplotlib
matplotlib.use('Agg')  # Non-interactive backend
import matplotlib.pyplot as plt

# Create a small test datapoint with shape (batch, channels, height, width)
# Using distinct dimensions: batch=2, channels=3, height=4, width=5
datapoint = np.random.rand(2, 3, 4, 5).astype(np.float32)

# Create a predicted array with the same shape
predicted = np.random.rand(2, 3, 4, 5).astype(np.float32)

ARGS = {
    'datapoint': datapoint,
    'predicted': predicted,
    'title_append': 'test suffix',
    'no_title': False,
    'show_plot': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (4, 13, 3) — shape (4, 13, 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

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 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.

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

Neural networks are emerging as a tool for scalable data-driven simulation of highdimensional dynamical systems, especially in settings where numerical methods are infeasible or computationally expensive. Notably, it has been shown that incorporating domain symmetries in deterministic neural simulators can substantially improve their accuracy, sample efficiency, and parameter efficiency. However, to incorporate symmetries in probabilistic neural simulators that can simulate stochastic phenomena, we need a model that produces equivariant distributions over trajectories, rather than equivariant function approximations. In this paper, we propose Equivariant Probabilistic Neural Simulation (EPNS), a framework for autoregressive probabilistic modeling of equivariant distributions over system evolutions. We use EPNS to design models for a stochastic n-body system and stochastic cellular dynamics. Our results show that EPNS considerably outperforms existing neural network-based methods for probabilistic simulation. More specifically, we demonstrate that incorporating equivariance in EPNS improves simulation quality, data efficiency, rollout stability, and uncertainty quantification. We conclude that EPNS is a promising method for efficient and effective data-driven probabilistic simulation in a diverse range of domains.

For agents

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

get_harvested_code_for_paper("2305.14286")
get_code_for_paper("2305.14286")
have("2305.14286")

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

curl https://syntology.ai/api/ran/2305.14286.json

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

[![Syntology run record](https://syntology.ai/api/ran/2305.14286.svg)](https://syntology.ai/paper/2305.14286)

Connect an agent — have() is free.