SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1807.10459 · 2018

IDTxl: The Information Dynamics Toolkit xl: a Python package for the efficient analysis of multivariate information dynamics in networks

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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 10 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/1807.10459.json · badge: /api/ran/1807.10459.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
dizcza/EmbedderSDR community (classifier label)
classifier basis not recorded on the link
10 ran · 0 no recorded run
FunctionStatusWhere it lives
calc_overlap Ran
ran on a synthesized input
dizcza/EmbedderSDR/experiment/kwta_inverse.py
code served (permissive licence) · get_code("d0492d5eb1772375")
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 two small batches of binary vectors
# batch_size=2, flattened_dim=4
vec1 = torch.tensor([[1, 0, 1, 0], [1, 1, 0, 1]], dtype=torch.float32)
vec2 = torch.tensor([[1, 1, 0, 0], [0, 1, 1, 1]], dtype=torch.float32)

# Reshape to have a batch dimension and spatial/feature dimensions
# This allows flatten(start_dim=1) to work properly
vec1 = vec1.reshape(2, 2, 2)  # batch_size=2, spatial dims=2x2
vec2 = vec2.reshape(2, 2, 2)  # batch_size=2, spatial dims=2x2

ARGS = {
    'vec1': vec1,
    'vec2': vec2,
}
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
densenet121 Ran
ran on a synthesized input
dizcza/EmbedderSDR/models/caltech.py
code served (permissive licence) · get_code("59ba360e21044bc8")
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 DenseNet. 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
import torchvision

torch.manual_seed(0)

# Create a minimal kwta module that has a forward method
class MinimalKWTA(nn.Module):
    def __init__(self, k=10):
        super().__init__()
        self.k = k
    
    def forward(self, x):
        return x

# Test case 1: without kwta
kwta_module = None

# Test case 2: with kwta (uncomment to use)
# kwta_module = MinimalKWTA(k=10)

ARGS = {
    'pretrained': False,
    'kwta': kwta_module
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DenseNet — a DenseNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DenseNet 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
get_kwta_threshold Ran
ran on a synthesized input
dizcza/EmbedderSDR/models/kwta.py
code served (permissive licence) · get_code("3c484fd885c2c829")
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, 1) as a later run recorded it. 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, 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
import math

torch.manual_seed(0)

class SparsityPredictor(nn.Module):
    def __init__(self, in_features, max_sparsity, min_sparsity=0.001):
        super().__init__()
        assert min_sparsity < max_sparsity
        if in_features is None or in_features == 'auto':
            self.linear = None
        else:
            self.linear = nn.Linear(in_features, out_features=1, bias=False)
        self.max_sparsity = max_sparsity
        self.min_sparsity = min_sparsity

    def forward(self, x):
        if self.linear is None:
            self.linear = nn.Linear(x.shape[1], out_features=1, bias=False)
        sparsity = self.linear(x).sigmoid() * self.max_sparsity
        return sparsity.squeeze()

    def extra_repr(self):
        extra = f'max_sparsity={self.max_sparsity}'
        if self.linear is None:
            extra = f"{extra}, linear='auto'"
        return extra

    def forward_threshold(self, x):
        assert x.ndimension() == 2, 'Input tensor is assumed to be flattened'
        sparsity = self.forward(x)
        scale = x.std(dim=1)
        loc = x.mean(dim=1)
        gaussian = torch.distributions.Normal(loc, scale)
        sparsity = sparsity.clamp_min(min=self.min_sparsity)
        threshold = gaussian.icdf(1.0 - sparsity)
        return threshold

def get_kwta_threshold(tensor, sparsity):
    unsqueeze_dim = [1] * (tensor.ndimension() - 1)
    tensor = tensor.flatten(start_dim=1)
    embedding_dim = tensor.shape[1]
    if embedding_dim < 2:
        raise ValueError(f'Embedding dimension {embedding_dim} should be >= 2')
    if isinstance(sparsity, SparsityPredictor):
        threshold = sparsity.forward_threshold(tensor)
    else:
        k_active = math.ceil(sparsity * embedding_dim)
        if k_active == embedding_dim:
            k_active -= 1
        topk = tensor.topk(k_active + 1, dim=1).values
        threshold = topk[:, [-2, -1]].mean(dim=1)
    threshold = threshold.view(-1, *unsqueeze_dim)
    return threshold

tensor = torch.randn(2, 4)
sparsity = 0.5

ARGS = {
    'tensor': tensor,
    'sparsity': sparsity
}
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, 1) — shape (2, 1) 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
make_feature_layers Ran
ran on a synthesized input
dizcza/EmbedderSDR/models/cifar/cifar.py
code served (permissive licence) · get_code("359797248803a427")
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 Sequential. 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)

# The function expects:
# - cfg: a list/iterable where each element is either 'M' (for MaxPool2d) or an integer/tuple
#        If tuple: (out_channels, padding)
#        If integer: out_channels (padding defaults to 1)
# - batch_norm: boolean flag

cfg = [
    64,           # Conv2d with 64 out_channels, padding=1
    (128, 1),     # Conv2d with 128 out_channels, padding=1
    'M',          # MaxPool2d
    256,          # Conv2d with 256 out_channels, padding=1
    (512, 0),     # Conv2d with 512 out_channels, padding=0
    'M',          # MaxPool2d
]

ARGS = {
    'cfg': cfg,
    'batch_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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
replace_relu Ran
ran on a synthesized input
dizcza/EmbedderSDR/utils/layers.py
code served (permissive licence) · get_code("6b4ebf0a4029a89d")
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 Sequential. 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
import copy

torch.manual_seed(0)

# Create a simple model with ReLU layers to test replace_relu
model = nn.Sequential(
    nn.Linear(4, 8),
    nn.ReLU(),
    nn.BatchNorm1d(8),
    nn.Linear(8, 3),
    nn.LeakyReLU(0.1),
    nn.ReLU6()
)

# Create a new ReLU activation function to replace with
new_relu = nn.ReLU()

# Define drop_layers as a tuple of layer types to replace with Identity
drop_layers = (nn.BatchNorm1d,)

ARGS = {
    "model": model,
    "new_relu": new_relu,
    "drop_layers": drop_layers
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
resnet18 Ran
ran on a synthesized input
dizcza/EmbedderSDR/models/caltech.py
code served (permissive licence) · get_code("c83c5836f49058cc")
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 ResNet. 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
import torchvision

torch.manual_seed(0)

# Create a minimal kwta module that accepts input and returns output
class MinimalKWTA(nn.Module):
    def __init__(self, out_features=512):
        super().__init__()
        self.out_features = out_features
    
    def forward(self, x):
        return x

# Test case 1: kwta=None (uses softmax classifier)
kwta_none = None

# Test case 2: kwta=module (uses kwta classifier)
kwta_module = MinimalKWTA(out_features=512)

ARGS = {
    'pretrained': False,
    'kwta': kwta_none
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResNet — a ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNet 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
squeezenet1_1 Ran
ran on a synthesized input
dizcza/EmbedderSDR/models/caltech.py
code served (permissive licence) · get_code("cf901cfe80efb8bc")
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 SqueezeNet. 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
import torchvision

torch.manual_seed(0)

# Define a minimal kwta module for testing
class MinimalKWTA(nn.Module):
    def forward(self, x):
        return x

# Create ARGS dictionary with all parameters for squeezenet1_1
ARGS = {
    'pretrained': False,  # Use False to avoid downloading pretrained weights
    'kwta': MinimalKWTA()  # Pass a minimal kwta module instance
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SqueezeNet — a SqueezeNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SqueezeNet 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
synthesize_log_softmax_data Ran
ran on a synthesized input
dizcza/EmbedderSDR/experiment/mutual_information.py
code served (permissive licence) · get_code("3c1416f5985217c4")
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 torch

torch.manual_seed(0)

ARGS = {
    "n_samples": 2,
    "n_classes": 4,
    "p_argmax": 0.95,
    "onehot": 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 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
torch_to_matplotlib Ran
ran on a synthesized input
dizcza/EmbedderSDR/experiment/kwta_inverse.py
code served (permissive licence) · get_code("49e970245ede8494")
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, 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 expects a tensor with a .shape attribute and supports
# .squeeze(dim=0), .transpose(0, 1), and .transpose(1, 2) operations.
# It handles two cases:
# 1. If shape[0] == 1: squeeze dimension 0
# 2. Otherwise: transpose dimensions 0->1 and 1->2

# Create a test case with shape[0] == 1 (single channel image)
image_single_channel = torch.randn(1, 4, 3)

ARGS = {
    "image": image_single_channel
}
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, 3) — shape (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
undo_normalization Ran
ran on a synthesized input
dizcza/EmbedderSDR/experiment/kwta_inverse.py
code served (permissive licence) · get_code("101d92d71a73b8e4")
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 torch.nn as nn

torch.manual_seed(0)

# Create normalized images - batch of 2, 4 channels/dims
images_normalized = [
    torch.tensor([0.1, 0.2, 0.3, 0.4]),
    torch.tensor([0.5, 0.6, 0.7, 0.8])
]

# Create a normalize_inverse function that denormalizes by reversing a normalization
# The function is called on each image and should return a tensor
def normalize_inverse(normalized_image):
    # Simple denormalization: multiply by 2 and add small offset
    denormalized = normalized_image * 2.0 + 0.1
    return denormalized

ARGS = {
    "images_normalized": images_normalized,
    "normalize_inverse": normalize_inverse
}
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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 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 10 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 Information Dynamics Toolkit xl (IDTxl) is a comprehensive software package for efficient inference of networks and their node dynamics from multivariate time series data using information theory. IDTxl provides functionality to estimate the following measures: 1) For network inference: multivariate transfer entropy (TE)/Granger causality (GC), multivariate mutual information (MI), bivariate TE/GC, bivariate MI 2) For analysis of node dynamics: active information storage (AIS), partial information decomposition (PID) IDTxl implements estimators for discrete and continuous data with parallel computing engines for both GPU and CPU platforms. Written for Python3.4.3+.

For agents

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

get_harvested_code_for_paper("1807.10459")
get_code_for_paper("1807.10459")
have("1807.10459")

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

curl https://syntology.ai/api/ran/1807.10459.json

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

[![Syntology run record](https://syntology.ai/api/ran/1807.10459.svg)](https://syntology.ai/paper/1807.10459)

Connect an agent — have() is free.