SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1810.03440 · 2018

Probabilistic Solutions To Ordinary Differential Equations As Non-Linear Bayesian Filtering: A New Perspective

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 10 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1810.03440.json · badge: /api/ran/1810.03440.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
mlysy/rodeo community (classifier label)
classifier basis not recorded on the link
10 ran · 0 no recorded run
FunctionStatusWhere it lives
add_sqrt Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/utils.py
code served (permissive licence) · get_code("d5951d6b51e09896")
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 ArrayImpl, ndim 2, dtype family float, shape (4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import jax.numpy as jnp
import numpy as np

# Create small square matrices for sqrt_A and sqrt_B
# Using dimension 4 as specified
n_dim = 4

# Create a simple positive semi-definite matrix and take its square root
# For sqrt_A: start with a random matrix and ensure it's positive semi-definite
np.random.seed(0)
A_temp = np.random.randn(n_dim, n_dim)
A = A_temp @ A_temp.T  # Ensure positive semi-definite
sqrt_A_np = np.linalg.cholesky(A)

# For sqrt_B: similar process
B_temp = np.random.randn(n_dim, n_dim)
B = B_temp @ B_temp.T  # Ensure positive semi-definite
sqrt_B_np = np.linalg.cholesky(B)

# Convert to jax arrays
sqrt_A = jnp.array(sqrt_A_np, dtype=jnp.float32)
sqrt_B = jnp.array(sqrt_B_np, dtype=jnp.float32)

ARGS = {
    "sqrt_A": sqrt_A,
    "sqrt_B": sqrt_B,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 2, dtype family float, shape (4, 4) — shape (4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ibm_init Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/prior/ibm.py
code served (permissive licence) · get_code("9bf758a66c464281")
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 4, 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 jax
import jax.numpy as jnp
import jax.scipy as jsp
import numpy as np

ARGS = {
    "dt": 0.1,
    "n_deriv": 3,
    "sigma": jnp.array([0.5, 1.0, 1.5, 2.0]),
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 4, 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
ibm_state Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/prior/ibm.py
code served (permissive licence) · get_code("6384560039110b11")
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 jax.numpy as jnp
import jax.scipy as jsp

ARGS = {
    "dt": 0.1,
    "q": 2,
    "sigma": 1.5,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — 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
indep_init Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/prior/indep_init.py
code served (permissive licence) · get_code("0f20bcd284871be1")
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 4, 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
import jax.numpy as jnp
import jax.scipy.linalg as jsl

# Create small block matrices for prior parameters
# prior_weight should be a tuple/list of matrices to be block diagonalized
block1_weight = jnp.array([[1.0, 0.5], [0.2, 0.9]])
block2_weight = jnp.array([[0.8, 0.1], [0.3, 0.7]])
prior_weight = (block1_weight, block2_weight)

# prior_var should be a tuple/list of variance matrices to be block diagonalized
block1_var = jnp.array([[0.5, 0.1], [0.1, 0.4]])
block2_var = jnp.array([[0.6, 0.05], [0.05, 0.5]])
prior_var = (block1_var, block2_var)

# Combine into prior_pars tuple
prior_pars = (prior_weight, prior_var)

ARGS = {
    "prior_pars": prior_pars
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 4, 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
interrogate_chkrebtii Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/interrogate.py
code served (permissive licence) · get_code("e7b5bf0434acb9a2")
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 jax
import jax.numpy as jnp
import jax.random as jr

# Set seed for reproducibility
key = jr.PRNGKey(0)

# Define small dimensions
n_block = 2  # number of blocks
n_bstate = 3  # dimension of state per block
n_weight = 2  # dimension of weight output

# Create ODE function that takes (state, t, **params) and returns predictions
def ode_fun(x_state, t, scale=1.0):
    """Simple ODE function: returns negative of scaled state"""
    return -scale * x_state

# Create ode_weight: shape (n_block, n_weight, n_bstate)
# This will be used in matrix multiplication: wm @ vsp @ wm.T
ode_weight = jnp.ones((n_block, n_weight, n_bstate))

# Create mean_state_pred: shape (n_block, n_bstate)
mean_state_pred = jnp.ones((n_block, n_bstate))

# Create var_state_pred: shape (n_block, n_bstate, n_bstate)
# Must be positive semi-definite for multivariate_normal
var_state_pred = jnp.array([
    jnp.eye(n_bstate) * 0.5,
    jnp.eye(n_bstate) * 0.5
])

# Time parameter
t = 0.0

# Kalman filter type
kalman_type = 'standard'

# Additional parameters for ode_fun
params = {'scale': 1.0}

# Construct ARGS dictionary with all function parameters
ARGS = {
    'key': key,
    'ode_fun': ode_fun,
    'ode_weight': ode_weight,
    't': t,
    'mean_state_pred': mean_state_pred,
    'var_state_pred': var_state_pred,
    'kalman_type': kalman_type,
    **params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — 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
interrogate_kramer Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/interrogate.py
code served (permissive licence) · get_code("ed22721f7952b81d")
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 jax
import jax.numpy as jnp

# Set up random seed for reproducibility
key = jax.random.PRNGKey(0)

# Define parameters based on the function signature
n_block = 2  # number of blocks
n_bmeas = 3  # measurement dimension per block
n_bstate = 4  # state dimension per block

# Create ode_weight tensor with shape (n_block, n_bmeas, n_bstate)
ode_weight = jnp.ones((n_block, n_bmeas, n_bstate))

# Create time parameter
t = 0.5

# Create mean_state_pred with shape (n_block, n_bstate)
mean_state_pred = jnp.array([[1.0, 2.0, 3.0, 4.0],
                              [5.0, 6.0, 7.0, 8.0]])

# Create var_state_pred with shape (n_block, n_bstate, n_bstate)
var_state_pred = jnp.ones((n_block, n_bstate, n_bstate))

# Define a simple ODE function that takes (state, t, **params)
def ode_fun(state, t, **params):
    """
    Simple ODE function.
    state has shape (n_block, n_bstate)
    returns array of same shape
    """
    return 0.1 * state

# Additional params dict (can be empty or contain ODE parameters)
params = {}

ARGS = {
    "key": key,
    "ode_fun": ode_fun,
    "ode_weight": ode_weight,
    "t": t,
    "mean_state_pred": mean_state_pred,
    "var_state_pred": var_state_pred,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — 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
interrogate_schober Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/interrogate.py
code served (permissive licence) · get_code("cde6cd9bff696a55")
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 jax
import jax.numpy as jnp

# Set up random key for JAX
key = jax.random.PRNGKey(0)

# Define a simple ODE function that takes (state, t, **params) and returns derivatives
def ode_fun(state, t, **params):
    """Simple ODE function: returns negative of state scaled by a parameter."""
    scale = params.get('scale', 1.0)
    return -scale * state

# Create ode_weight tensor with shape (n_block, n_bmeas, n_bmeas)
# Using distinct dimensions: 2, 3, 3
n_block = 2
n_bmeas = 3
ode_weight = jnp.ones((n_block, n_bmeas, n_bmeas))

# Time parameter
t = 0.5

# Mean state prediction: shape (n_block, state_dim)
# Using state_dim = 4 to be distinct from n_bmeas = 3
mean_state_pred = jnp.ones((n_block, 4))

# Variance state prediction: shape (n_block, state_dim, state_dim)
var_state_pred = jnp.eye(4).reshape(1, 4, 4) * 0.1
var_state_pred = jnp.tile(var_state_pred, (n_block, 1, 1))

# Additional parameters for the ODE function
params = {'scale': 1.5}

ARGS = {
    'key': key,
    'ode_fun': ode_fun,
    'ode_weight': ode_weight,
    't': t,
    'mean_state_pred': mean_state_pred,
    'var_state_pred': var_state_pred,
}

# Merge params into ARGS
ARGS.update(params)
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — 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
multivariate_normal_logpdf Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/utils.py
code served (permissive licence) · get_code("1cdc64790cf0fa88")
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 ArrayImpl, 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 jax.numpy as jnp
import numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Dimension of the multivariate normal
p = 4

# Create observation vector x
x = jnp.array(np.random.randn(p), dtype=jnp.float32)

# Create mean vector
mean = jnp.array(np.random.randn(p), dtype=jnp.float32)

# Create a symmetric positive definite covariance matrix
# Generate random matrix and multiply by its transpose to ensure positive definiteness
A = np.random.randn(p, p)
cov_np = A @ A.T + np.eye(p)  # Add identity to ensure positive definiteness
cov = jnp.array(cov_np, dtype=jnp.float32)

ARGS = {
    "x": x,
    "mean": mean,
    "cov": cov,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
mvncond Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/utils.py
code served (permissive licence) · get_code("377de7eb841e8bad")
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 jax.numpy as jnp
import numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create a small multivariate normal setup
# mu: mean vector of dimension 2*n_dim
# Sigma: covariance matrix of dimension (2*n_dim, 2*n_dim)
# icond: boolean array indicating which dimensions to condition on

n_dim = 2
total_dim = 2 * n_dim  # 4 dimensions total

# Create a mean vector
mu = jnp.array([1.0, 2.0, 3.0, 4.0])

# Create a positive definite covariance matrix
# Start with a random matrix and make it positive definite
base_matrix = np.random.randn(total_dim, total_dim)
Sigma = np.dot(base_matrix, base_matrix.T) + np.eye(total_dim)
Sigma = jnp.array(Sigma, dtype=jnp.float32)

# Create a boolean conditioning array
# True means we condition on that dimension, False means we want to predict it
icond = jnp.array([True, False, True, False])

ARGS = {
    "mu": mu,
    "Sigma": Sigma,
    "icond": icond
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — 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
predict Ran
ran on a synthesized input
mlysy/rodeo/src/rodeo/kalmantv/standard.py
code served (permissive licence) · get_code("38225b69bc66f85b")
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 jax.numpy as jnp
import numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Define state dimension
n_state = 4

# Create valid arguments for the predict function
mean_state_past = jnp.array([1.0, 2.0, 3.0, 4.0])
var_state_past = jnp.array([
    [1.0, 0.1, 0.0, 0.0],
    [0.1, 1.0, 0.1, 0.0],
    [0.0, 0.1, 1.0, 0.1],
    [0.0, 0.0, 0.1, 1.0]
])

mean_state = jnp.array([0.5, 0.5, 0.5, 0.5])
wgt_state = jnp.array([
    [0.9, 0.0, 0.0, 0.0],
    [0.0, 0.9, 0.0, 0.0],
    [0.0, 0.0, 0.9, 0.0],
    [0.0, 0.0, 0.0, 0.9]
])

var_state = jnp.array([
    [0.1, 0.0, 0.0, 0.0],
    [0.0, 0.1, 0.0, 0.0],
    [0.0, 0.0, 0.1, 0.0],
    [0.0, 0.0, 0.0, 0.1]
])

ARGS = {
    'mean_state_past': mean_state_past,
    'var_state_past': var_state_past,
    'mean_state': mean_state,
    'wgt_state': wgt_state,
    'var_state': var_state,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — 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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 10 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We formulate probabilistic numerical approximations to solutions of ordinary differential equations (ODEs) as problems in Gaussian process (GP) regression with non-linear measurement functions. This is achieved by defining the measurement sequence to consist of the observations of the difference between the derivative of the GP and the vector field evaluated at the GP---which are all identically zero at the solution of the ODE. When the GP has a state-space representation, the problem can be reduced to a non-linear Bayesian filtering problem and all widely-used approximations to the Bayesian filtering and smoothing problems become applicable. Furthermore, all previous GP-based ODE solvers that are formulated in terms of generating synthetic measurements of the gradient field come out as specific approximations. Based on the non-linear Bayesian filtering problem posed in this paper, we develop novel Gaussian solvers for which we establish favourable stability properties. Additionally, non-Gaussian approximations to the filtering problem are derived by the particle filter approach. The resulting solvers are compared with other probabilistic solvers in illustrative experiments.

For agents

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

get_harvested_code_for_paper("1810.03440")
get_code_for_paper("1810.03440")
have("1810.03440")

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

curl https://syntology.ai/api/ran/1810.03440.json

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

[![Syntology run record](https://syntology.ai/api/ran/1810.03440.svg)](https://syntology.ai/paper/1810.03440)

Connect an agent — have() is free.