SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2607.23554 · 2026

Neonatal Hypoxic-ischaemic Encephalopathy Classification from the EEG and HRV Signals Using a Conformer based Masked Autoencoder

Shuwen Yu, Gordon Lightbody, William Marnane, Geraldine Boylan

arXiv · PDF · Open in the Atlas

Code that ran

6 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures of our instrument; 5 no recorded run), measured 2026-11-22; 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/2607.23554.json · badge: /api/ran/2607.23554.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
syu-kylin/MAEConformer community
Papers with Code archive
6 ran · 5 no recorded run
FunctionStatusWhere it lives
ConvolutionModule Ran
ran on a synthesized input · constructed an object; no result was computed
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("90b2c49faa4d8f06")
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 an instance of its own class, ConvolutionModule: the run constructed the object and did not call it, so no result was computed. 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)

# Define the GRN class (required by ConvolutionModule)
class GRN(nn.Module):
    """ GRN (Global Response Normalization) layer
    """

    def __init__(self, dim):
        super().__init__()
        self.gamma = nn.Parameter(torch.zeros(1, dim, 1))
        self.beta = nn.Parameter(torch.zeros(1, dim, 1))

    def forward(self, x):
        """
        Args:
            x (torch.Tensor): with shape of (batch, d_model, seqLence).
        Returns:
            torch.Tensor: with shape of (batch, d_model, seqLence).
        """
        Gx = torch.norm(x, p=2, dim=(2,), keepdim=True)
        Nx = Gx / (Gx.mean(dim=1, keepdim=True) + 1e-06)
        out = self.gamma * (x * Nx) + self.beta + x
        return out

# Define the ConvolutionModule class
class ConvolutionModule(nn.Module):
    """
    Confermer convolution module.
    Args: 
        input_dim (int): input dimension.
        depthwise_kernel_size (int): kernel size of depthwise convolution layer.
        dropout (float, optional): dropout probability. (Default: 0.0)
        bias (bool, optional): whether to add bias term to each convolution layer. (DefaultL False)
        use_group_norm (bool, optional): use group normal (instance normal) or batchnorm. (Default False (batchnorm))
    """

    def __init__(self, d_model: int, depthwise_kernel_size: int, dropout: float=0.0, bias: bool=False, use_group_norm: bool=False) -> None:
        super().__init__()
        if (depthwise_kernel_size - 1) % 2 != 0:
            raise ValueError('depthwise_kernel_size must be odd to achieve same padding.')
        self.layernorm = nn.LayerNorm(d_model)
        self.sequential = nn.Sequential(nn.Conv1d(d_model, 2 * d_model, kernel_size=1, stride=1, padding=0, bias=bias), nn.GLU(dim=1), nn.Conv1d(d_model, d_model, depthwise_kernel_size, stride=1, padding=(depthwise_kernel_size - 1) // 2, groups=d_model, bias=bias), GRN(d_model), nn.SiLU(), nn.Conv1d(d_model, d_model, kernel_size=1, stride=1, padding=0, bias=bias), nn.Dropout(dropout))

    def forward(self, input: torch.Tensor) -> torch.Tensor:
        """
        Args:
            input (torch.Tensor): with shape of (batch, seqLence, d_model).
        Returns:
            torch.Tensor: with shape of (batch, seqLence, d_model).
        """
        x = self.layernorm(input)
        x = x.transpose(1, 2)
        x = self.sequential(x)
        return x.transpose(1, 2)

# Create test input
batch_size = 2
seq_length = 5
d_model = 4

test_input = torch.randn(batch_size, seq_length, d_model)

# Define ARGS dict with all parameters
ARGS = {
    'd_model': 4,
    'depthwise_kernel_size': 3,
    'dropout': 0.1,
    'bias': False,
    'use_group_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 ConvolutionModule — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ConvolutionModule 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
GRN Ran
ran on a synthesized input
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("a59bbbc89bd654a6")
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, 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

class GRN(nn.Module):
    """ GRN (Global Response Normalization) layer
    """

    def __init__(self, dim):
        super().__init__()
        self.gamma = nn.Parameter(torch.zeros(1, dim, 1))
        self.beta = nn.Parameter(torch.zeros(1, dim, 1))

    def forward(self, x):
        """
        Args:
            x (torch.Tensor): with shape of (batch, d_model, seqLence).
        Returns:
            torch.Tensor: with shape of (batch, d_model, seqLence).
        """
        Gx = torch.norm(x, p=2, dim=(2,), keepdim=True)
        Nx = Gx / (Gx.mean(dim=1, keepdim=True) + 1e-06)
        out = self.gamma * (x * Nx) + self.beta + x
        return out


INIT_ARGS = {
    'dim': 4
}

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

if __name__ == '__main__':
    grn = GRN(**INIT_ARGS)
    output = grn(**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, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
MRSTFTLoss Ran
ran on a synthesized input · constructed an object; no result was computed
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("9347103050baf48b")
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 an instance of its own class, MRSTFTLoss: the run constructed the object and did not call it, so no result was computed. 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)

# Create small test tensors
batch_size = 2
seq_len = 3
patch_size = 4

# pred: (B, seq_len, patch_size) reconstructed signal
pred = torch.randn(batch_size, seq_len, patch_size)

# target: (B, seq_len, patch_size) ground truth signal
target = torch.randn(batch_size, seq_len, patch_size)

# mask: (B, seq_len) binary mask indicating masked patches (1=masked)
mask = torch.randint(0, 2, (batch_size, seq_len))

ARGS = {
    "win_lengths": (64, 128, 256),
    "hop_ratio": 0.5,
    "mag_weight": 1.0,
    "cpx_weight": 0.5,
    "mask_aware": True,
    "min_frame_weight": 0.0,
    "center": False,
    "eps": 1e-08,
}
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 MRSTFTLoss — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MRSTFTLoss 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
MSELoss Ran
ran on a synthesized input
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("77ecfa4c52991e52")
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 0, 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

INIT_ARGS = {}

# Create sample tensors for the forward method
B = 2
seq_len = 5
patch_size = 8

pred = torch.randn(B, seq_len, patch_size)
target = torch.randn(B, seq_len, patch_size)
mask = torch.tensor([[1, 0, 1, 0, 1], [0, 1, 1, 0, 1]], dtype=torch.float32)

ARGS = {
    'pred': pred,
    'target': target,
    'mask': mask
}
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 0, 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)
PositionEncoding Ran
ran on a synthesized input · constructed an object; no result was computed
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("5bce251634e2bc7f")
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 an instance of its own class, PositionEncoding: the run constructed the object and did not call it, so no result was computed. 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 PositionEncoding class is instantiated with d_model and max_len
# d_model: dimension of the model (embedding dimension)
# max_len: maximum sequence length for which to precompute positional encodings

d_model = 8
max_len = 16

ARGS = {
    "d_model": d_model,
    "max_len": max_len,
}
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 PositionEncoding — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PositionEncoding 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
RandomMask Ran
ran on a synthesized input · constructed an object; no result was computed
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("4a90ff4f913f3c66")
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 an instance of its own class, RandomMask: the run constructed the object and did not call it, so no result was computed. 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)

class RandomMask:

    def __init__(self, mask_ratio):
        self.mask_ratio = mask_ratio

    def __call__(self, x):
        """
        Args:
            x (torch.Tensor): with shape of (N, L, D) where N is batch size, L is sequence length, D is feature dimension.
        Returns:
            x_unmasked (torch.Tensor): with shape of (N, L * (1 - mask_ratio), D) where the masked patches are removed.
            mask (torch.Tensor): with shape of (N, L) where 0 means the patch is kept and 1 means the patch is masked.
            ids_restore (torch.Tensor): with shape of (N, L) where the original indexes of the patches are restored.
        """
        N, L, D = x.shape
        len_keep = int(L * (1 - self.mask_ratio))
        noise = torch.rand(N, L, device=x.device)
        ids_shuffle = torch.argsort(noise, dim=1)
        ids_restore = torch.argsort(ids_shuffle, dim=1)
        ids_keep = ids_shuffle[:, :len_keep]
        x_unmasked = torch.gather(x, dim=1, index=ids_keep.unsqueeze(-1).repeat(1, 1, D))
        return (x_unmasked, ids_restore)

# Construct valid arguments
mask_ratio = 0.3

# Create a sample input tensor with shape (N, L, D)
# N=2 (batch size), L=8 (sequence length), D=4 (feature dimension)
x = torch.randn(2, 8, 4)

ARGS = {
    "mask_ratio": mask_ratio
}
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 RandomMask — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a RandomMask 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
ConformerBlock Not yet run
unverified (harvested, no recorded run)
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("6aa7677871103422")
How it ran

no invocation recorded.

ConformerEncoder Not yet run
unverified (harvested, no recorded run)
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("d325b089742f0886")
How it ran

no invocation recorded.

MAEConformer Not yet run
unverified (harvested, no recorded run)
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("2e22a987854638d6")
How it ran

no invocation recorded.

XformerDecoder Not yet run
unverified (harvested, no recorded run)
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("ed852e7b4a55261c")
How it ran

no invocation recorded.

init_weights Not yet run
unverified (harvested, no recorded run)
syu-kylin/MAEConformer/model/MAEConformer.py
code served (permissive licence) · get_code("52febdee98eabcbd")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 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 11 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

In this paper, we propose the MAEConformer, a novel self-supervised learning framework that combines the Conformer architecture with the Masked Autoencoder (MAE) paradigm for large-scale representation learning from unlabelled electroencephalography (EEG) and heart rate variability (HRV) signals. By integrating convolutional operations with Transformer-based self-attention, MAE-Conformer effectively captures both local temporal patterns and long-range contextual dependencies in physiological time series. To enhance reconstruction fidelity and representation quality, a multiresolution short-time Fourier transform (MR-STFT) loss is incorporated alongside the reconstruction objective, enabling the model to jointly learn temporal and spectral characteristics across multiple scales. Modality-specific EEG and HRV MAEConformer models were pretrained on 6,030 h and 4,868 h of unlabelled recordings, respectively, and subsequently transferred to expert-annotated downstream tasks. Experimental results demonstrate that the learned representations provide strong transferability and data efficiency. In EEG-based hypoxic-ischemic encephalopathy (HIE) severity classification, the pretrained MAE-EEG model achieved test AUCs of 97.19% and 96.56% for binary and four-class classification tasks, respectively, outperforming a range of state-of-the-art supervised and self-supervised baselines. On the HRV-based HIE severity classification task, MAE-HRV achieved a test AUC of 82.42%, surpassing both self-supervised Transformer-based and supervised convolutional baselines. These findings demonstrate the effectiveness of MAEConformer for learning robust and transferable representations across multiple physiological modalities. The code is publicly available at https://github.com/syu-kylin/MAEConformer.

For agents

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

get_harvested_code_for_paper("2607.23554")
get_code_for_paper("2607.23554")
have("2607.23554")

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

curl https://syntology.ai/api/ran/2607.23554.json

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

[![Syntology run record](https://syntology.ai/api/ran/2607.23554.svg)](https://syntology.ai/paper/2607.23554)

Connect an agent — have() is free.