SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.01915 · ICML · 2024

Robust Inverse Graphics via Probabilistic Inference

Matthew Hoffman, Rif Saurous, Brian Patton, Tuan Le, Pavel Sountsov, Ben Lee

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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; 1 honoured the contract we drafted, 0 violated it and 0 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/2402.01915.json · badge: /api/ran/2402.01915.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
tensorflow/probability canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
FunctionStatusWhere it lives
convert_data_format Ran
ran; our contract draft was wrong, not the code
tensorflow/probability/tensorflow_probability/python/layers/conv_variational.py
code served (permissive licence) · get_code("61fb5425d07a822d")
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 str. 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

# Test case 1: channels_last with 3D input
ARGS = {
    'data_format': 'channels_last',
    'ndim': 3
}

# Alternative test cases to verify different paths:
# ARGS = {
#     'data_format': 'channels_last',
#     'ndim': 4
# }

# ARGS = {
#     'data_format': 'channels_last',
#     'ndim': 5
# }

# ARGS = {
#     'data_format': 'channels_first',
#     'ndim': 3
# }

# ARGS = {
#     'data_format': 'channels_first',
#     'ndim': 4
# }

# ARGS = {
#     'data_format': 'channels_first',
#     'ndim': 5
# }
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
normalize_data_format Ran
ran; our contract draft was wrong, not the code
tensorflow/probability/tensorflow_probability/python/layers/conv_variational.py
code served (permissive licence) · get_code("1dc93d89e545ab22")
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 str. 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

# Examine the function: normalize_data_format takes a single parameter 'value'
# It calls value.lower() and checks if the result is in {'channels_first', 'channels_last'}
# So 'value' needs to be a string-like object with a lower() method

ARGS = {
    'value': 'channels_first'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
normalize_tuple Ran
ran, honoured the contract we drafted
tensorflow/probability/tensorflow_probability/python/layers/conv_variational.py
code served (permissive licence) · get_code("f2557115172a2cae")
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 1, dtype family int; 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

def normalize_tuple(value, n, name):
    """Transforms a single integer or iterable of integers into an integer tuple.

  Args:
    value: The value to validate and convert. Could an int, or any iterable
      of ints.
    n: The size of the tuple to be returned.
    name: The name of the argument being validated, e.g. "strides" or
      "kernel_size". This is only used to format error messages.

  Returns:
    A tuple of n integers.

  Raises:
    ValueError: If something else than an int/long or iterable thereof was
      passed.
  """
    if isinstance(value, int):
        return (value,) * n
    else:
        try:
            value_tuple = tuple(value)
        except TypeError:
            raise ValueError(f'Argument `{name}` must be a tuple of {str(n)} integers. Received: {str(value)}') from None
        if len(value_tuple) != n:
            raise ValueError(f'Argument `{name}` must be a tuple of {str(n)} integers. Received: {str(value)}')
        for single_value in value_tuple:
            try:
                int(single_value)
            except (ValueError, TypeError):
                raise ValueError(f'Argument `{name}` must be a tuple of {str(n)} integers. Received: {str(value)} including element {str(single_value)} of type {str(type(single_value))}') from None
        return value_tuple

torch.manual_seed(0)

ARGS = {
    "value": 2,
    "n": 3,
    "name": "kernel_size"
}
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, ndim 1, dtype family int — 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

How do we infer a 3D scene from a single image in the presence of corruptions like rain, snow or fog? Straightforward domain randomization relies on knowing the family of corruptions ahead of time. Here, we propose a Bayesian approachdubbed robust inverse graphics (RIG)-that relies on a strong scene prior and an uninformative uniform corruption prior, making it applicable to a wide range of corruptions. Given a single image, RIG performs posterior inference jointly over the scene and the corruption. We demonstrate this idea by training a neural radiance field (NeRF) scene prior and using a secondary NeRF to represent the corruptions over which we place an uninformative prior. RIG, trained only on clean data, outperforms depth estimators and alternative NeRF approaches that perform point estimation instead of full inference. The results hold for a number of scene prior architectures based on normalizing flows and diffusion models. For the latter, we develop reconstruction-guidance with auxiliary latents (ReGAL)-a diffusion conditioning algorithm that is applicable in the presence of auxiliary latent variables such as the corruption. RIG demonstrates how scene priors can be used beyond generation tasks.

For agents

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

get_harvested_code_for_paper("2402.01915")
get_code_for_paper("2402.01915")
have("2402.01915")

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

curl https://syntology.ai/api/ran/2402.01915.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.01915.svg)](https://syntology.ai/paper/2402.01915)

Connect an agent — have() is free.