SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1609.09869 · 2016

Structured Inference Networks for Nonlinear State Space Models

arXiv · PDF · Open in the Atlas

Code that ran

13 of 13 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; 2 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/1609.09869.json · badge: /api/ran/1609.09869.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
clinicalml/structuredinference canonical (classifier label)
classifier basis: census
5 ran · 0 no recorded run
yjlolo/pytorch-deep-markov-model community (classifier label)
classifier basis not recorded on the link
8 ran · 0 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran; our contract draft was wrong, not the code
yjlolo/pytorch-deep-markov-model/model/metric.py
pointer only (licence: MIT); the stored copy get_code serves, from lingbai-kong/causalformer, is pointer only (licence: GPL-3.0) · get_code("5dcb4148be3fcbaf")
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 lingbai-kong/causalformer, and GPL-3.0, as GitHub detects it on that repository, is not in the permissive set Syntology serves inline. 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. A run is recorded.

dmm_loss Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/model/loss.py
code served (permissive licence) · get_code("c6893c2d184dfa80")
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
import torch.nn as nn

torch.manual_seed(0)

# Create tiny tensors with distinct dimensions
batch_size = 2
seq_length = 3
feature_dim = 4

x = torch.randn(batch_size, seq_length, feature_dim)
x_hat = torch.randn(batch_size, seq_length, feature_dim)
mu1 = torch.randn(batch_size, seq_length, feature_dim)
logvar1 = torch.randn(batch_size, seq_length, feature_dim)
mu2 = torch.randn(batch_size, seq_length, feature_dim)
logvar2 = torch.randn(batch_size, seq_length, feature_dim)

# Create a mask tensor with shape compatible with kl_fr and nll_fr
# After mean(dim=-1), they have shape (batch_size, seq_length)
# After view(-1), they have shape (batch_size * seq_length,)
# So mask should have shape (batch_size, seq_length) or be broadcastable
mask = torch.ones(batch_size, seq_length)

kl_annealing_factor = 0.5

ARGS = {
    "x": x,
    "x_hat": x_hat,
    "mu1": mu1,
    "logvar1": logvar1,
    "mu2": mu2,
    "logvar2": logvar2,
    "kl_annealing_factor": kl_annealing_factor,
    "mask": mask,
}
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
evaluateBound Ran
ran on a synthesized input
clinicalml/structuredinference/stinfmodel_fast/evaluate.py
code served (permissive licence) · get_code("b1fbeb47b9bcc52e")
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 float64, 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 numpy as np

class MockDKF:
    def __init__(self):
        self.dataset = None
        self.mask = None
    
    def resetDataset(self, dataset, mask):
        self.dataset = dataset
        self.mask = mask
    
    def evaluate(self, idx):
        # Return a vector of ELBO values for the given indices
        # Shape should be (len(idx),)
        return np.random.randn(len(idx))
    
    def _p(self, msg):
        # Mock print function
        pass

# Create small test data
np.random.seed(0)
N = 10  # number of sequences
T = 5   # time steps per sequence
feature_dim = 4

# Dataset: (N, T, feature_dim)
dataset = np.random.randn(N, T, feature_dim).astype(np.float32)

# Mask: (N, T) - binary mask indicating valid frames
mask = np.ones((N, T), dtype=np.float32)
mask[0, 4] = 0  # mask out one frame to have variation
mask[1, :3] = 0

# Create DKF model
dkf = MockDKF()

# Define ARGS
ARGS = {
    'dkf': dkf,
    'dataset': dataset,
    'mask': mask,
    'batch_size': 3,
    'S': 2,
    'normalization': 'frame',
    'additional': {}
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
generateData Ran
ran on a synthesized input
clinicalml/structuredinference/baselines/filters.py
code served (permissive licence) · get_code("43d35013b8f92339")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 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 numpy as np

def init_mu_dummy():
    return 0.0

def init_cov_dummy():
    return 1.0

def trans_fxn_dummy(z):
    return z * 0.9

def trans_cov_dummy():
    return 0.5

def obs_fxn_dummy(z):
    return z + 0.1

def obs_cov_dummy():
    return 0.2

params = {
    'dataset_A': {
        'init_mu': init_mu_dummy(),
        'init_cov': init_cov_dummy(),
        'trans_fxn': trans_fxn_dummy,
        'trans_cov': trans_cov_dummy(),
        'obs_fxn': obs_fxn_dummy,
        'obs_cov': obs_cov_dummy(),
    }
}

ARGS = {
    'N': 2,
    'T': 4,
    'params': params,
    'dname': 'dataset_A',
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_mini_batch_mask Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/data_loader/seq_util.py
code served (permissive licence) · get_code("44f52e0dc7d1314c")
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)

# Create input tensor (batch_size=2, max_sequence_length=4, input_dim=3)
x = torch.randn(2, 4, 3)

# Create sequence lengths tensor (batch_size=2)
# Each element indicates the actual length of sequences in the batch
seq_lengths = torch.tensor([2, 3], dtype=torch.long)

ARGS = {
    "x": x,
    "seq_lengths": seq_lengths,
}
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
impSamplingNLL Ran
ran on a synthesized input
clinicalml/structuredinference/stinfmodel_fast/evaluate.py
code served (permissive licence) · get_code("ad9f362bb16235ec")
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 float32, 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 numpy as np
import torch
import torch.nn as nn

torch.manual_seed(0)
np.random.seed(0)

class MockDKF(nn.Module):
    """Minimal mock DKF model with required methods"""
    def __init__(self):
        super().__init__()
        self.dataset = None
        self.mask = None
    
    def resetDataset(self, dataset, mask):
        """Store dataset and mask"""
        self.dataset = dataset
        self.mask = mask
    
    def likelihood(self, idx):
        """Return log-likelihood values for given indices"""
        # Return shape (batch_size, 1) for each call
        batch_size = len(idx)
        return np.random.randn(batch_size, 1).astype(np.float32)
    
    def meanSumExp(self, x, axis=1):
        """Compute log-sum-exp mean over axis"""
        # Numerical stability: log(mean(exp(x)))
        max_x = np.max(x, axis=axis, keepdims=True)
        exp_x = np.exp(x - max_x)
        mean_exp = np.mean(exp_x, axis=axis, keepdims=True)
        result = np.log(mean_exp) + max_x
        return result.squeeze(axis=axis)
    
    def _p(self, msg):
        """Print message"""
        print(msg)

# Create mock objects
dkf = MockDKF()

# Create small dataset: 6 sequences, each with 4 timesteps
dataset = np.random.randn(6, 4).astype(np.float32)

# Create mask: indicate which frames to use (all frames valid)
mask = np.ones((6, 4), dtype=bool)

batch_size = 2
S = 2
normalization = 'frame'

ARGS = {
    'dkf': dkf,
    'dataset': dataset,
    'mask': mask,
    'batch_size': batch_size,
    'S': S,
    'normalization': normalization
}
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 float32, 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
infer Ran
ran on a synthesized input
clinicalml/structuredinference/stinfmodel_fast/evaluate.py
code served (permissive licence) · get_code("f93223114e53959c")
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
import numpy as np
import torch.nn as nn

torch.manual_seed(0)

class SimpleRecognitionNet(nn.Module):
    """Minimal DKF-like model with required methods"""
    def __init__(self, dim_observations, dim_latent):
        super().__init__()
        self.params = {'dim_observations': dim_observations}
        self.dim_latent = dim_latent
        self.current_dataset = None
        self.current_mask = None
        
        # Simple recognition network layers
        self.fc1 = nn.Linear(dim_observations, 8)
        self.fc_mu = nn.Linear(8, dim_latent)
        self.fc_logcov = nn.Linear(8, dim_latent)
    
    def resetDataset(self, dataset, mask, quiet=False):
        """Store dataset and mask for inference"""
        self.current_dataset = dataset
        self.current_mask = mask
    
    def posterior_inference(self, idx):
        """Return z, mu, logcov tensors"""
        batch_size = len(idx)
        seq_len = self.current_dataset.shape[1]
        
        # Process data through recognition network
        data = self.current_dataset[idx]  # shape: (batch, seq_len, dim_obs)
        flat_data = data.reshape(-1, self.params['dim_observations'])
        
        h = torch.relu(self.fc1(flat_data))
        mu = self.fc_mu(h)
        logcov = self.fc_logcov(h)
        
        # Reshape to 3D: (batch, seq_len, dim_latent)
        mu = mu.reshape(batch_size, seq_len, self.dim_latent)
        logcov = logcov.reshape(batch_size, seq_len, self.dim_latent)
        
        # Sample z from posterior
        std = torch.exp(0.5 * logcov)
        z = mu + std * torch.randn_like(mu)
        
        return z, mu, logcov

# Create model
dim_obs = 5
dim_latent = 3
dkf = SimpleRecognitionNet(dim_observations=dim_obs, dim_latent=dim_latent)

# Create dataset: (batch_size=2, seq_len=4, dim_obs=5)
dataset = torch.randn(2, 4, dim_obs)

# Create mask: (batch_size=2, seq_len=4)
mask = torch.ones(2, 4, dtype=torch.bool)

ARGS = {
    'dkf': dkf,
    'dataset': dataset,
    'mask': mask
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — 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
kl_div Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/model/loss.py
code served (permissive licence) · get_code("8c4257c4dbe9e6de")
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)

# Create small tensors for mu1 and logvar1
# Shape: (batch_size=2, latent_dim=4)
mu1 = torch.randn(2, 4)
logvar1 = torch.randn(2, 4)

# mu2 and logvar2 are optional and default to zeros/zeros_like
# We'll provide explicit values to test the full function
mu2 = torch.randn(2, 4)
logvar2 = torch.randn(2, 4)

ARGS = {
    "mu1": mu1,
    "logvar1": logvar1,
    "mu2": mu2,
    "logvar2": logvar2,
}
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
nll_loss Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/model/loss.py
code served (permissive licence) · get_code("b5351f56bbfd3e22")
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
import torch.nn as nn

torch.manual_seed(0)

# Construct valid arguments for nll_loss
# The function expects:
# - x_hat: 3D tensor (logits for BCEWithLogitsLoss)
# - x: 3D tensor with same shape as x_hat (targets, should be in [0, 1] for BCE)

batch_size = 2
seq_len = 4
feature_dim = 3

x_hat = torch.randn(batch_size, seq_len, feature_dim)
x = torch.rand(batch_size, seq_len, feature_dim)

ARGS = {
    'x_hat': x_hat,
    '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 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
pad_and_reverse Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/data_loader/seq_util.py
code served (permissive licence) · get_code("9d6c12fb45bba0cb")
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, 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 packed sequence as input to pad_and_reverse
batch_size = 2
max_time = 5
input_dim = 4

# Create raw RNN output (unpacked)
rnn_output_unpacked = torch.randn(batch_size, max_time, input_dim)

# Create sequence lengths (actual lengths of sequences before padding)
seq_lengths = torch.tensor([3, 4], dtype=torch.long)

# Pack the sequence
packed = nn.utils.rnn.pack_padded_sequence(
    rnn_output_unpacked,
    seq_lengths,
    batch_first=True,
    enforce_sorted=False
)

ARGS = {
    "rnn_output": packed,
    "seq_lengths": seq_lengths,
}
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, 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: none by the harness; any seed is in the setup block
reverse_sequence Ran
ran on a synthesized input
yjlolo/pytorch-deep-markov-model/data_loader/seq_util.py
code served (permissive licence) · get_code("95a4c36e248fe204")
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, 5, 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)

# x: tensor (b, T_max, input_dim)
# seq_lengths: tensor (b, )
# The function reverses sequences along the time axis (dimension 1)
# based on the actual length of each sequence in the batch

batch_size = 2
T_max = 5
input_dim = 4

# Create sample input tensor
x = torch.randn(batch_size, T_max, input_dim)

# Create sequence lengths tensor
# Each element indicates the actual length of that sequence
seq_lengths = torch.tensor([3, 4], dtype=torch.long)

ARGS = {
    "x": x,
    "seq_lengths": seq_lengths,
}
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, 5, 4) — shape (2, 5, 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
sampleGaussian Ran
ran on a synthesized input
clinicalml/structuredinference/baselines/filters.py
code served (permissive licence) · get_code("117016d0884d1f3b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 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 numpy as np

np.random.seed(0)

# mu must be a numpy array
mu = np.array([1.0, 2.0, 3.0, 4.0])

# cov can be scalar or same shape as mu
cov = 0.5

ARGS = {
    "mu": mu,
    "cov": cov
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 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
top_k_acc Ran
ran; our contract draft was wrong, not the code
yjlolo/pytorch-deep-markov-model/model/metric.py
pointer only (licence: MIT); the stored copy get_code serves, from lingbai-kong/causalformer, is pointer only (licence: GPL-3.0) · get_code("61cc9667c37ecd27")
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 lingbai-kong/causalformer, and GPL-3.0, as GitHub detects it on that repository, is not in the permissive set Syntology serves inline. 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. A run is recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 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.

FunctionAlso present inEarliest-datedOthers named
accuracy 5 other papers
5 later
1907.00865 (2019-07)
Radial Bayesian Neural Networks: Beyond Discrete Support In Large-Scale Bayesian Deep Learning
2111.04578 (2021-11)
Improved Regularization and Robustness for Fine-tuning in Neural Networks
2206.02659 (2022-06)
Robust Fine-Tuning of Deep Neural Networks with Hessian-based Generalization Guarantees
and 2 more not named here
top_k_acc 4 other papers
4 later
2111.04578 (2021-11)
Improved Regularization and Robustness for Fine-tuning in Neural Networks
2206.02659 (2022-06)
Robust Fine-Tuning of Deep Neural Networks with Hessian-based Generalization Guarantees
2312.10890 (2023-12)
Low-latency Space-time Supersampling for Real-time Rendering
and 1 more not named here

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

Gaussian state space models have been used for decades as generative models of sequential data. They admit an intuitive probabilistic interpretation, have a simple functional form, and enjoy widespread adoption. We introduce a unified algorithm to efficiently learn a broad class of linear and non-linear state space models, including variants where the emission and transition distributions are modeled by deep neural networks. Our learning algorithm simultaneously learns a compiled inference network and the generative model, leveraging a structured variational approximation parameterized by recurrent neural networks to mimic the posterior distribution. We apply the learning algorithm to both synthetic and real-world datasets, demonstrating its scalability and versatility. We find that using the structured approximation to the posterior results in models with significantly higher held-out likelihood.

For agents

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

get_harvested_code_for_paper("1609.09869")
get_code_for_paper("1609.09869")
have("1609.09869")

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

curl https://syntology.ai/api/ran/1609.09869.json

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

[![Syntology run record](https://syntology.ai/api/ran/1609.09869.svg)](https://syntology.ai/paper/1609.09869)

Connect an agent — have() is free.