SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.01371 · 2024

eXponential FAmily Dynamical Systems (XFADS): Large-scale nonlinear Gaussian state-space modeling

arXiv · PDF · Open in the Atlas

Code that ran

8 of 14 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 6 failures of our instrument; 6 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/2403.01371.json · badge: /api/ran/2403.01371.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
catniplab/xfads canonical (classifier label)
classifier basis: census
8 ran · 6 no recorded run
FunctionStatusWhere it lives
bmv Ran
ran, honoured the contract we drafted
catniplab/xfads/xfads/linalg_utils.py
code served (permissive licence) · get_code("66fab7e673f1a01c")
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

torch.manual_seed(0)

# A is a batch of matrices: shape (batch, n, n)
# x is a batch of vectors: shape (batch, n)
# The function computes batched matrix-vector product

A = torch.randn(2, 4, 4)  # batch=2, matrix is 4x4
x = torch.randn(2, 4)     # batch=2, vector is 4-dimensional

ARGS = {
    "A": A,
    "x": x,
}
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 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
bop Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/linalg_utils.py
code served (permissive licence) · get_code("5d15ccc22a9c48ab")
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 3, dtype family float, shape (2, 4, 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 torch

torch.manual_seed(0)

# The function bop takes two tensors and performs an einsum operation
# It expects tensors with compatible shapes for the einsum '...i, ...j -> ...ij'
# This means both tensors should have at least 1 dimension, with the last dimension
# being treated as 'i' and 'j' respectively

# Create two small tensors with distinct sizes
# x1 shape: (2, 4) - batch size 2, feature dim 4
# x2 shape: (2, 3) - batch size 2, feature dim 3
x1 = torch.randn(2, 4)
x2 = torch.randn(2, 3)

ARGS = {
    'x1': x1,
    'x2': x2
}
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 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 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
build_gru_dynamics_function Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/utils.py
code served (permissive licence) · get_code("b35422bc6fba813f")
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 DynamicsGRU. 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

torch.manual_seed(0)

ARGS = {
    'dim_input': 4,
    'dim_hidden': 8,
    'd_type': torch.float32,
    'device': 'cpu',
    'use_layer_norm': False,
}
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 DynamicsGRU — a DynamicsGRU output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DynamicsGRU 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
estimate_poisson_rate_bias Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/prob_utils.py
code served (permissive licence) · get_code("3d034c33d26e2408")
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 1, dtype family float, shape (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)

# Test case 1: Direct tensor input
y_tensor = torch.randn(2, 3, 4)  # batch=2, time=3, features=4
time_delta_1 = 0.5

# Test case 2: DataLoader input
class SimpleDataset:
    def __init__(self, num_samples=4):
        self.num_samples = num_samples
    
    def __iter__(self):
        for i in range(self.num_samples):
            y_sample = torch.randn(2, 3, 4)  # batch=2, time=3, features=4
            x_dummy = torch.randn(2, 5)
            yield (y_sample, y_sample)
    
    def __len__(self):
        return self.num_samples

dataloader = SimpleDataset(num_samples=2)
time_delta_2 = 0.5

# Using tensor for primary test
ARGS = {
    'y': y_tensor,
    'time_delta': time_delta_1
}
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 1, dtype family float, shape (4,) — shape (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
memory_cleanup Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/decorators.py
code served (permissive licence) · get_code("204798d70e07dfc6")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 functools
import gc
import torch.nn as nn

torch.manual_seed(0)

# Create a simple test function to decorate
def dummy_function(x, y):
    """A simple function that performs some computation."""
    return x + y

# Decorate the function
decorated_function = lambda x, y: dummy_function(x, y)

# Create test arguments
test_tensor_1 = torch.randn(2, 4)
test_tensor_2 = torch.randn(2, 4)

ARGS = {
    "func": dummy_function
}
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
show_memory_stats Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/decorators.py
code served (permissive licence) · get_code("f113aa7c337f472a")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 functools

def show_memory_stats(func):
    @functools.wraps(func)
    def wrapper(*args, **kwargs):
        if torch.cuda.is_available():
            print(f'\nBefore {func.__name__}\nAllocated: {torch.cuda.memory_allocated(0) / 1000000000.0} GB   Reserved: {torch.cuda.memory_reserved(0) / 1000000000.0} GB   Free: {(torch.cuda.get_device_properties(0).total_memory - torch.cuda.memory_reserved(0)) / 1000000000.0} GB\n')
            result = func(*args, **kwargs)
            print(f'\nAfter {func.__name__}\nAllocated: {torch.cuda.memory_allocated(0) / 1000000000.0} GB   Reserved: {torch.cuda.memory_reserved(0) / 1000000000.0} GB   Free: {(torch.cuda.get_device_properties(0).total_memory - torch.cuda.memory_reserved(0)) / 1000000000.0} GB\n')
        else:
            print('You are using the show_memory_stats function decorator but, no CUDA detected, function executed casually.')
            result = func(*args, **kwargs)
        return result
    return wrapper

torch.manual_seed(0)

def simple_function(x, y):
    """A simple test function to decorate"""
    return x + y

decorated_func = show_memory_stats(simple_function)

ARGS = {
    'func': decorated_func
}
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
softplus_inv Ran
ran, honoured the contract we drafted
catniplab/xfads/xfads/utils.py
code served (permissive licence) · get_code("61afa93ae14d2d28")
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 softplus_inv takes a single parameter x
# which can be either a torch.Tensor or a numpy array
# Let's test both cases by creating valid inputs

# Case 1: torch.Tensor input
x_tensor = torch.randn(2, 4) + 2.0  # Add 2.0 to ensure values are positive and in valid range
x_tensor = torch.abs(x_tensor) + 0.5  # Ensure positive values for softplus_inv

# Case 2: numpy array input
x_numpy = np.random.randn(2, 4) + 2.0
x_numpy = np.abs(x_numpy) + 0.5  # Ensure positive values

# We'll use the torch.Tensor version as the primary test case
ARGS = {
    "x": x_tensor
}
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
triangular_inverse Ran
ran; our contract draft was wrong, not the code
catniplab/xfads/xfads/linalg_utils.py
code served (permissive licence) · get_code("47f1e66bbd89dca4")
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 (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
import torch.nn as nn

torch.manual_seed(0)

# Create a small triangular matrix (lower triangular for testing)
# Size: 4x4 to match the eye() call in the function
A = torch.tensor([
    [2.0, 0.0, 0.0, 0.0],
    [1.0, 3.0, 0.0, 0.0],
    [0.5, 0.2, 4.0, 0.0],
    [0.1, 0.3, 0.4, 5.0]
], dtype=torch.float32)

ARGS = {
    "A": A,
    "upper": False
}
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 2, dtype family float, shape (4, 4) — shape (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
animate_reaching_evolution Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/plot_utils.py
code served (permissive licence) · get_code("c17790ba37349e87")
How it ran

no invocation recorded.

animate_reaching_evolution_mo Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/plot_utils.py
code served (permissive licence) · get_code("137195894580df0c")
How it ran

no invocation recorded.

apply_memory_cleanup Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/decorators.py
code served (permissive licence) · get_code("52b222b6a45a0eed")
How it ran

no invocation recorded.

bits_per_spike Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/prob_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from arsedler9/lfads-torch, is pointer only (licence: NOASSERTION) · get_code("e44e57b951a765b2")
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 arsedler9/lfads-torch, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

plot_single_trial_reaching_latents Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/plot_utils.py
code served (permissive licence) · get_code("aeb58bd43522d4b8")
How it ran

no invocation recorded.

spike_resample_fn Not yet run
unverified (harvested, no recorded run)
catniplab/xfads/xfads/utils.py
code served (permissive licence) · get_code("ad5739732afb4b50")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 13 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
bits_per_spike 2 other papers
1 dated earlier, 1 later
2109.04463 (2021-09)
Neural Latents Benchmark '21: Evaluating latent variable models of neural population activity
2512.00239 (2025-12)
Self-Supervised Dynamical System Representations for Physiological Time-Series

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

State-space graphical models and the variational autoencoder framework provide a principled apparatus for learning dynamical systems from data. State-of-the-art probabilistic approaches are often able to scale to large problems at the cost of flexibility of the variational posterior or expressivity of the dynamics model. However, those consolidations can be detrimental if the ultimate goal is to learn a generative model capable of explaining the spatiotemporal structure of the data and making accurate forecasts. We introduce a low-rank structured variational autoencoding framework for nonlinear Gaussian state-space graphical models capable of capturing dense covariance structures that are important for learning dynamical systems with predictive capabilities. Our inference algorithm exploits the covariance structures that arise naturally from sample based approximate Gaussian message passing and low-rank amortized posterior updates -- effectively performing approximate variational smoothing with time complexity scaling linearly in the state dimensionality. In comparisons with other deep state-space model architectures our approach consistently demonstrates the ability to learn a more predictive generative model. Furthermore, when applied to neural physiological recordings, our approach is able to learn a dynamical system capable of forecasting population spiking and behavioral correlates from a small portion of single trials.

For agents

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

get_harvested_code_for_paper("2403.01371")
get_code_for_paper("2403.01371")
have("2403.01371")

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

curl https://syntology.ai/api/ran/2403.01371.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.01371.svg)](https://syntology.ai/paper/2403.01371)

Connect an agent — have() is free.