SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2111.08679 · NeurIPS · 2021

Automatically detecting anomalous exoplanet transits

Benjamin Miller, Christoph Hönes, Ana Heras, Bernard Foing

arXiv · PDF · Open in the Atlas

Code that ran

12 of 25 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 11 have no contract verdict; 1 failure of our instrument; 13 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/2111.08679.json · badge: /api/ran/2111.08679.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
christophhoenes/anomalousexoplanettransits canonical (classifier label)
classifier basis: classifier_2025
12 ran · 13 no recorded run
FunctionStatusWhere it lives
ConvLayer Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("cee46964326d3513")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 8, 16) 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
import numpy as np
import warnings

def select_activation_function(activation):
    if activation in ['ReLU', 'relu']:
        return nn.ReLU()
    elif activation.startswith('LeakyReLU') or activation.startswith('leaky_relu'):
        try:
            act_strip = activation.strip('LeakyReLU')
            act_strip = act_strip.strip('leaky_relu')
            act_strip = act_strip.strip('_')
            return nn.LeakyReLU(negative_slope=float(act_strip))
        except ValueError:
            return nn.LeakyReLU()
    elif activation in ['PReLU', 'prelu']:
        return nn.PReLU()
    elif activation in ['GELU', 'gelu']:
        return nn.GELU()
    elif activation in ['ELU', 'elu']:
        return nn.ELU()
    elif activation in ['SELU', 'selu']:
        return nn.SELU()
    elif activation in ['Tanh', 'tanh']:
        return nn.Tanh()
    elif activation in ['Sigmoid', 'sigmoid']:
        return nn.Sigmoid()
    elif activation in ['Softplus', 'softplus']:
        return nn.Softplus
    else:
        warnings.warn('Activation function "{}" is unknown! No activation will be used.'.format(activation), RuntimeWarning)
        return None

class ConvLayer(nn.Module):

    def __init__(self, in_channels=16, out_channels=16, kernel_size=5, stride=1, padding=None, padding_mode='zeros', dilation=1, activation='ReLU', batch_norm=True, transpose=False, bias=None):
        super(ConvLayer, self).__init__()
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.kernel_size = kernel_size
        self.stride = stride
        self.padding = padding
        self.dilation = dilation
        self.activation = activation
        self.batch_norm = batch_norm
        if padding is None:
            padding_amount = int(np.ceil(dilation * (kernel_size - 1) / 2))
        else:
            padding_amount = padding
        if transpose:
            modules = [nn.ConvTranspose1d(in_channels, out_channels, kernel_size, stride=stride, padding=padding_amount, padding_mode=padding_mode, output_padding=stride - 1 if padding_amount > 0 else 0, dilation=dilation, bias=not batch_norm if bias is None else bias)]
        else:
            modules = [nn.Conv1d(in_channels, out_channels, kernel_size, stride=stride, padding=padding_amount, padding_mode=padding_mode, dilation=dilation, bias=not batch_norm if bias is None else bias)]
        if batch_norm:
            modules.append(nn.BatchNorm1d(out_channels))
        if activation is not None:
            modules.append(select_activation_function(activation))
        self._components = nn.Sequential(*modules)

    def forward(self, x):
        return self._components(x)


INIT_ARGS = {
    'in_channels': 4,
    'out_channels': 8,
    'kernel_size': 3,
    'stride': 1,
    'padding': None,
    'padding_mode': 'zeros',
    'dilation': 1,
    'activation': 'ReLU',
    'batch_norm': True,
    'transpose': False,
    'bias': None,
}

ARGS = {
    'x': torch.randn(2, 4, 16),
}

if __name__ == '__main__':
    layer = ConvLayer(**INIT_ARGS)
    output = layer(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 3, dtype family float, shape (2, 8, 16) — shape (2, 8, 16) 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
GlobalAvgPool1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("73780839e6b08582")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, 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

# INIT_ARGS for GlobalAvgPool1d.__init__
INIT_ARGS = {
    'map_size': None,
    'channel_dim': -1
}

# ARGS for GlobalAvgPool1d.forward
# Create a 3D tensor (batch, channels, sequence_length)
# Using distinct dimensions: batch=2, channels=5, sequence_length=7
x = torch.randn(2, 5, 7)

ARGS = {
    'x': x
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
GlobalMaxPool1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("124991c1258ee4f2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 1) 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

class GlobalMaxPool1d(nn.Module):

    def __init__(self, map_size=None, channel_dim=-1):
        super(GlobalMaxPool1d, self).__init__()
        self.map_size = map_size
        self.channel_dim = channel_dim

    def forward(self, x):
        return x.max(dim=self.channel_dim, keepdim=True).values


INIT_ARGS = {
    'map_size': None,
    'channel_dim': -1
}

ARGS = {
    'x': torch.randn(2, 4, 3)
}

if __name__ == '__main__':
    model = GlobalMaxPool1d(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 1) — shape (2, 4, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
LinearLayer Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("cb07fa1cfe45efb6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 2, dtype family float, shape (2, 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
import torch.nn as nn
import warnings

def select_activation_function(activation):
    if activation in ['ReLU', 'relu']:
        return nn.ReLU()
    elif activation.startswith('LeakyReLU') or activation.startswith('leaky_relu'):
        try:
            act_strip = activation.strip('LeakyReLU')
            act_strip = act_strip.strip('leaky_relu')
            act_strip = act_strip.strip('_')
            return nn.LeakyReLU(negative_slope=float(act_strip))
        except ValueError:
            return nn.LeakyReLU()
    elif activation in ['PReLU', 'prelu']:
        return nn.PReLU()
    elif activation in ['GELU', 'gelu']:
        return nn.GELU()
    elif activation in ['ELU', 'elu']:
        return nn.ELU()
    elif activation in ['SELU', 'selu']:
        return nn.SELU()
    elif activation in ['Tanh', 'tanh']:
        return nn.Tanh()
    elif activation in ['Sigmoid', 'sigmoid']:
        return nn.Sigmoid()
    elif activation in ['Softplus', 'softplus']:
        return nn.Softplus()
    else:
        warnings.warn('Activation function "{}" is unknown! No activation will be used.'.format(activation), RuntimeWarning)
        return None

class LinearLayer(nn.Module):

    def __init__(self, in_channels=16, out_channels=16, dropout=0.0, activation=None, batch_norm=False):
        super(LinearLayer, self).__init__()
        self.activation = activation
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.dropout = dropout
        self.batch_norm = batch_norm
        modules = [nn.Linear(in_channels, out_channels)]
        if self.activation is not None:
            modules.append(select_activation_function(activation))
        if batch_norm:
            modules.append(nn.BatchNorm1d(out_channels))
        if dropout > 0.0:
            modules.append(nn.Dropout(p=self.dropout))
        self._components = nn.Sequential(*modules)

    def forward(self, x):
        return self._components(x)


INIT_ARGS = {
    'in_channels': 4,
    'out_channels': 3,
    'dropout': 0.1,
    'activation': 'ReLU',
    'batch_norm': True
}

ARGS = {
    'x': torch.randn(2, 4)
}

layer = LinearLayer(**INIT_ARGS)
output = layer(**ARGS)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 3) — shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
MeanPool1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("e80ba618fb269491")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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
import torch.nn as nn
import torch.nn.functional as F

class MeanPool1d(nn.Module):

    def __init__(self, kernel_size, stride, padding=0, return_indices=True):
        super(MeanPool1d, self).__init__()
        self.kernel_size = kernel_size
        self.stride = stride
        self.return_indices = return_indices
        self.padding = padding

    def forward(self, x):
        if self.return_indices:
            return (F.avg_pool1d(x, self.kernel_size, stride=self.stride, padding=self.padding), None)
        else:
            F.avg_pool1d(x, self.kernel_size, stride=self.stride, padding=self.padding)


INIT_ARGS = {
    'kernel_size': 3,
    'stride': 2,
    'padding': 1,
    'return_indices': True
}

ARGS = {
    'x': torch.randn(2, 4, 8)
}

if __name__ == '__main__':
    model = MeanPool1d(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output type: {type(output)}")
    print(f"Output shape: {output[0].shape if isinstance(output, tuple) else output.shape}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
PoolingLayer Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("3862b8e4d2bc7b10")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, 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 torch.nn as nn
import warnings
import torch.nn.functional as F

class GlobalAvgPool1d(nn.Module):

    def __init__(self, map_size=None, channel_dim=-1):
        super(GlobalAvgPool1d, self).__init__()
        self.map_size = map_size
        self.channel_dim = channel_dim

    def forward(self, x):
        return x.mean(dim=self.channel_dim, keepdim=True)

class GlobalMaxPool1d(nn.Module):

    def __init__(self, map_size=None, channel_dim=-1):
        super(GlobalMaxPool1d, self).__init__()
        self.map_size = map_size
        self.channel_dim = channel_dim

    def forward(self, x):
        return x.max(dim=self.channel_dim, keepdim=True).values

class MeanPool1d(nn.Module):

    def __init__(self, kernel_size, stride, padding=0, return_indices=True):
        super(MeanPool1d, self).__init__()
        self.kernel_size = kernel_size
        self.stride = stride
        self.return_indices = return_indices
        self.padding = padding

    def forward(self, x):
        if self.return_indices:
            return (F.avg_pool1d(x, self.kernel_size, stride=self.stride, padding=self.padding), None)
        else:
            F.avg_pool1d(x, self.kernel_size, stride=self.stride, padding=self.padding)

class StridePool1d(nn.Module):

    def __init__(self, pool_kernel, pool_stride, channels, padding=0, activation='ReLU', batch_norm=False):
        super(StridePool1d, self).__init__()
        self.layer = nn.Conv1d(channels, channels, kernel_size=pool_kernel, stride=pool_stride, padding=padding)
        self.activation = select_activation_function(activation)
        self.batch_norm = nn.BatchNorm1d(channels) if batch_norm else None

    def forward(self, x):
        out = self.layer(x) if self.batch_norm is None else self.batch_norm(self.layer(x))
        return out if self.activation is None else self.activation(out)

def select_activation_function(activation):
    if activation in ['ReLU', 'relu']:
        return nn.ReLU()
    elif activation.startswith('LeakyReLU') or activation.startswith('leaky_relu'):
        try:
            act_strip = activation.strip('LeakyReLU')
            act_strip = act_strip.strip('leaky_relu')
            act_strip = act_strip.strip('_')
            return nn.LeakyReLU(negative_slope=float(act_strip))
        except ValueError:
            return nn.LeakyReLU()
    elif activation in ['PReLU', 'prelu']:
        return nn.PReLU()
    elif activation in ['GELU', 'gelu']:
        return nn.GELU()
    elif activation in ['ELU', 'elu']:
        return nn.ELU()
    elif activation in ['SELU', 'selu']:
        return nn.SELU()
    elif activation in ['Tanh', 'tanh']:
        return nn.Tanh()
    elif activation in ['Sigmoid', 'sigmoid']:
        return nn.Sigmoid()
    elif activation in ['Softplus', 'softplus']:
        return nn.Softplus
    else:
        warnings.warn('Activation function "{}" is unknown! No activation will be used.'.format(activation), RuntimeWarning)
        return None

class PoolingLayer(nn.Module):

    def __init__(self, out_channels=None, pool_type='stride', pool_kernel=2, pool_stride=2, pool_padding=None, map_size=None):
        super(PoolingLayer, self).__init__()
        self.out_channels = out_channels
        self.pool_type = pool_type
        self.pool_kernel = pool_kernel
        self.pool_stride = pool_stride
        self.pool_padding = pool_padding
        self.map_size = map_size
        self._indices = None
        if pool_padding == None:
            pool_padding = (pool_kernel - 1) // 2
        if pool_type == 'max_indices':
            self.pooling = nn.MaxPool1d(pool_kernel, pool_stride, padding=pool_padding, return_indices=True)
        elif pool_type == 'max':
            self.pooling = nn.MaxPool1d(pool_kernel, pool_stride, padding=pool_padding, return_indices=False)
        elif pool_type == 'mean':
            self.pooling = MeanPool1d(pool_kernel, pool_stride, padding=pool_padding, return_indices=False)
        elif pool_type == 'global_max':
            self.pooling = GlobalMaxPool1d(map_size)
        elif pool_type == 'global_avg':
            self.pooling = GlobalAvgPool1d(map_size)
        elif pool_type == 'stride':
            assert out_channels is not None, "Passing the number of channels of the previous layer is required for pool_type 'stride'!"
            self.pooling = StridePool1d(pool_kernel, pool_stride, out_channels, padding=pool_padding)
        else:
            raise RuntimeWarning('Unknown Pooling function {} encountered!'.format(self.pooling))

    @property
    def indices(self):
        return self._indices

    def forward(self, x):
        if self.pool_type == 'max_indices':
            out, indices = self.pooling(x)
            self._indices = indices
            return out
        return self.pooling(x)


INIT_ARGS = {
    'out_channels': 4,
    'pool_type': 'stride',
    'pool_kernel': 2,
    'pool_stride': 2,
    'pool_padding': None,
    'map_size': None,
}

ARGS = {
    'x': torch.randn(2, 4, 8),
}

if __name__ == '__main__':
    model = PoolingLayer(**INIT_ARGS)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
StridePool1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("238ca52eff8cac7a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 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
import warnings

def select_activation_function(activation):
    if activation in ['ReLU', 'relu']:
        return nn.ReLU()
    elif activation.startswith('LeakyReLU') or activation.startswith('leaky_relu'):
        try:
            act_strip = activation.strip('LeakyReLU')
            act_strip = act_strip.strip('leaky_relu')
            act_strip = act_strip.strip('_')
            return nn.LeakyReLU(negative_slope=float(act_strip))
        except ValueError:
            return nn.LeakyReLU()
    elif activation in ['PReLU', 'prelu']:
        return nn.PReLU()
    elif activation in ['GELU', 'gelu']:
        return nn.GELU()
    elif activation in ['ELU', 'elu']:
        return nn.ELU()
    elif activation in ['SELU', 'selu']:
        return nn.SELU()
    elif activation in ['Tanh', 'tanh']:
        return nn.Tanh()
    elif activation in ['Sigmoid', 'sigmoid']:
        return nn.Sigmoid()
    elif activation in ['Softplus', 'softplus']:
        return nn.Softplus()
    else:
        warnings.warn('Activation function "{}" is unknown! No activation will be used.'.format(activation), RuntimeWarning)
        return None

class StridePool1d(nn.Module):

    def __init__(self, pool_kernel, pool_stride, channels, padding=0, activation='ReLU', batch_norm=False):
        super(StridePool1d, self).__init__()
        self.layer = nn.Conv1d(channels, channels, kernel_size=pool_kernel, stride=pool_stride, padding=padding)
        self.activation = select_activation_function(activation)
        self.batch_norm = nn.BatchNorm1d(channels) if batch_norm else None

    def forward(self, x):
        out = self.layer(x) if self.batch_norm is None else self.batch_norm(self.layer(x))
        return out if self.activation is None else self.activation(out)


INIT_ARGS = {
    'pool_kernel': 3,
    'pool_stride': 2,
    'channels': 4,
    'padding': 1,
    'activation': 'ReLU',
    'batch_norm': False
}

ARGS = {
    'x': torch.randn(2, 4, 8)
}

if __name__ == '__main__':
    model = StridePool1d(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print("Success!")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 4) — shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
StrideUnPool1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("fb3ad5d0baba37c6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 17) 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
import warnings

def select_activation_function(activation):
    if activation in ['ReLU', 'relu']:
        return nn.ReLU()
    elif activation.startswith('LeakyReLU') or activation.startswith('leaky_relu'):
        try:
            act_strip = activation.strip('LeakyReLU')
            act_strip = act_strip.strip('leaky_relu')
            act_strip = act_strip.strip('_')
            return nn.LeakyReLU(negative_slope=float(act_strip))
        except ValueError:
            return nn.LeakyReLU()
    elif activation in ['PReLU', 'prelu']:
        return nn.PReLU()
    elif activation in ['GELU', 'gelu']:
        return nn.GELU()
    elif activation in ['ELU', 'elu']:
        return nn.ELU()
    elif activation in ['SELU', 'selu']:
        return nn.SELU()
    elif activation in ['Tanh', 'tanh']:
        return nn.Tanh()
    elif activation in ['Sigmoid', 'sigmoid']:
        return nn.Sigmoid()
    elif activation in ['Softplus', 'softplus']:
        return nn.Softplus
    else:
        warnings.warn('Activation function "{}" is unknown! No activation will be used.'.format(activation), RuntimeWarning)
        return None

class StrideUnPool1d(nn.Module):

    def __init__(self, pool_kernel, pool_stride, channels, padding=0, activation='ReLU', batch_norm=False):
        super(StrideUnPool1d, self).__init__()
        self.layer = nn.ConvTranspose1d(channels, channels, kernel_size=pool_kernel, stride=pool_stride, padding=padding)
        self.activation = select_activation_function(activation)
        self.batch_norm = nn.BatchNorm1d(channels) if batch_norm else None

    def forward(self, x):
        out = self.layer(x) if self.batch_norm is None else self.batch_norm(self.layer(x))
        return out if self.activation is None else self.activation(out)


INIT_ARGS = {
    'pool_kernel': 3,
    'pool_stride': 2,
    'channels': 4,
    'padding': 0,
    'activation': 'ReLU',
    'batch_norm': False
}

ARGS = {
    'x': torch.randn(2, 4, 8)
}

if __name__ == '__main__':
    model = StrideUnPool1d(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 17) — shape (2, 4, 17) 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
UnFlatten Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("07a296880eef7407")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 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

class UnFlatten(nn.Module):

    def __init__(self, channels):
        super(UnFlatten, self).__init__()
        self.channels = channels

    def forward(self, x):
        return x.unflatten(1, (self.channels, x.size()[1] // self.channels))


INIT_ARGS = {
    'channels': 4
}

ARGS = {
    'x': torch.randn(2, 16)
}

if __name__ == '__main__':
    model = UnFlatten(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print(f"Output:\n{output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 4) — shape (2, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Upsample1d Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("2d764b73a39b8e5e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 6) 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

class Upsample1d(torch.nn.Module):

    def __init__(self, scale_factor):
        super(Upsample1d, self).__init__()
        self._scale_factor = scale_factor
        self._module = nn.UpsamplingNearest2d(scale_factor=scale_factor)

    @property
    def scale_factor(self):
        return self._scale_factor

    def forward(self, x):
        out = torch.unsqueeze(x, 3)
        out = self._module(out)
        return out[:, :, :, 0]


INIT_ARGS = {
    'scale_factor': 2
}

ARGS = {
    'x': torch.randn(2, 4, 3)
}

if __name__ == '__main__':
    model = Upsample1d(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print(f"Scale factor: {model.scale_factor}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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, 6) — shape (2, 4, 6) 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
VerticalFitter Ran
ran on a synthesized input
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("502c4c43de0bc288")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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
import torch.nn as nn
import torch.nn.functional as F

class VerticalFitter(nn.Module):

    def __init__(self, size):
        super(VerticalFitter, self).__init__()
        self.influence_trend1 = torch.linspace(1.0, 0.0, size).unsqueeze(dim=0).unsqueeze(dim=0)
        self.influence_trend2 = torch.ones(size) - self.influence_trend1

    def _apply(self, fn):
        super(VerticalFitter, self)._apply(fn)
        self.influence_trend1 = fn(self.influence_trend1)
        self.influence_trend2 = fn(self.influence_trend2)
        return self

    def forward(self, data, model, scale, trend):
        oot_model = model.max(dim=-1, keepdims=True).values
        model -= oot_model
        model *= 1.0 - scale.unsqueeze(dim=1).unsqueeze(dim=1)
        model += trend[:, 0].unsqueeze(dim=1).unsqueeze(dim=1) * self.influence_trend1 + trend[:, 1].unsqueeze(dim=1).unsqueeze(dim=1) * self.influence_trend2
        vertical_loss = F.mse_loss(model, data)
        return (model.detach(), vertical_loss)


INIT_ARGS = {
    'size': 5
}

ARGS = {
    'data': torch.randn(2, 3, 5),
    'model': torch.randn(2, 3, 5),
    'scale': torch.randn(2),
    'trend': torch.randn(2, 2)
}

if __name__ == '__main__':
    fitter = VerticalFitter(**INIT_ARGS)
    output_model, loss = fitter(**ARGS)
    print(f"Output model shape: {output_model.shape}")
    print(f"Loss: {loss.item()}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
select_activation_function Ran
ran; our contract draft was wrong, not the code
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("68a933eee54ea1c9")
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) and returned ReLU. 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 warnings

torch.manual_seed(0)

# The function takes a single parameter: activation (a string)
# It checks the activation string against various patterns and returns
# the corresponding nn.Module or None

ARGS = {
    'activation': 'ReLU'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ReLU — a ReLU output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ReLU 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
CustomVAE Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("fd6c571d2dd20aa2")
How it ran

no invocation recorded.

MaxUnPool1d Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("da8a1627fadb7154")
How it ran

no invocation recorded.

MeanUnool1d Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("86d470da61f23a5f")
How it ran

no invocation recorded.

ResBlock Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("325d2e47b1c6ab8b")
How it ran

no invocation recorded.

TransitShapeVAE Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("32e152287c3fa2e3")
How it ran

no invocation recorded.

UnPoolingLayer Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("c576df2e5f57b644")
How it ran

no invocation recorded.

VAE Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("ae4700404a606f78")
How it ran

no invocation recorded.

create_custom_network Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("aa45df8e57bc564e")
How it ran

no invocation recorded.

create_custom_network Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("bb380f0f47bc9c7c")
How it ran

no invocation recorded.

local_outlier_factor Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/anomaly_detection.py
code served (permissive licence) · get_code("c678e7ec5c49e802")
How it ran

no invocation recorded.

mahalanobis_distance Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/anomaly_detection.py
code served (permissive licence) · get_code("88de2f51668310c9")
How it ran

no invocation recorded. this sample carries JSON test cases instead of a setup block; get_reference_implementation serves them with the level they back.

one_class_svm Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/anomaly_detection.py
code served (permissive licence) · get_code("1e9a2a20ec63f538")
How it ran

no invocation recorded.

select_activation_function Not yet run
unverified (harvested, no recorded run)
christophhoenes/anomalousexoplanettransits/architectures.py
code served (permissive licence) · get_code("de4eb295a57a97e1")
How it ran

no invocation recorded. this sample carries JSON test cases instead of a setup block; get_reference_implementation serves them with the level they back.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 23 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 23 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

Raw light curve data from exoplanet transits is too complex to naively apply traditional outlier detection methods. We propose an architecture which estimates a latent representation of both the main transit and residual deviations with a pair of variational autoencoders. We show, using two fabricated datasets, that our latent representations of anomalous transit residuals are significantly more amenable to outlier detection than raw data or the latent representation of a traditional variational autoencoder. We then apply our method to real exoplanet transit data. Our study is the first which automatically identifies anomalous exoplanet transit light curves. We additionally release three first-of-their-kind datasets to enable further research.

For agents

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

get_harvested_code_for_paper("2111.08679")
get_code_for_paper("2111.08679")
have("2111.08679")

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

curl https://syntology.ai/api/ran/2111.08679.json

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

[![Syntology run record](https://syntology.ai/api/ran/2111.08679.svg)](https://syntology.ai/paper/2111.08679)

Connect an agent — have() is free.