SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.16908 · 2020

MDP Homomorphic Networks: Group Symmetries in Reinforcement Learning

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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 6 have no contract verdict; 1 failure 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/2006.16908.json · badge: /api/ran/2006.16908.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
ElisevanderPol/mdp-homomorphic-networks canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
ElisevanderPol/symmetrizer community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
closed_group Ran
ran on a synthesized input
ElisevanderPol/symmetrizer/symmetrizer/groups/ops.py
code served (permissive licence) · get_code("8a531fe4a9ab6edd")
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 bool, 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 torch
import numpy as np

torch.manual_seed(0)

# Create a list of small 2D matrices (elements) that form a closed group
# Using simple 2x2 identity and rotation matrices
element1 = torch.tensor([[1.0, 0.0], [0.0, 1.0]])  # Identity
element2 = torch.tensor([[0.0, -1.0], [1.0, 0.0]])  # 90 degree rotation
element3 = torch.tensor([[-1.0, 0.0], [0.0, -1.0]])  # 180 degree rotation
element4 = torch.tensor([[0.0, 1.0], [-1.0, 0.0]])  # 270 degree rotation

elements = [element1, element2, element3, element4]

ARGS = {
    "elements": elements
}
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 bool, 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
conv2d_output_shape Ran
ran on a synthesized input
ElisevanderPol/mdp-homomorphic-networks/rlpyt/models/utils.py
code served (permissive licence) · get_code("fc2d9d7f350a539e")
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

torch.manual_seed(0)

# Analyze the function:
# - h: input height (int)
# - w: input width (int)
# - kernel_size: int or tuple of ints (default 1)
# - stride: int or tuple of ints (default 1)
# - padding: int or tuple of ints (default 0)
# - dilation: int (default 1)

# The function performs a conv2d output shape calculation
# All parameters are used in arithmetic operations
# kernel_size, stride, padding can be int or tuple
# We'll test with basic valid values

ARGS = {
    "h": 32,
    "w": 64,
    "kernel_size": 3,
    "stride": 1,
    "padding": 1,
    "dilation": 1,
}
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
get_basis Ran
ran on a synthesized input
ElisevanderPol/symmetrizer/symmetrizer/ops/ops.py
code served (permissive licence) · get_code("4f1f3a4c76b650d9")
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 numpy as np
import torch

torch.manual_seed(0)

class MockGroup:
    """Mock symmetry group with required methods and attributes"""
    def __init__(self):
        self.parameters = ['param1', 'param2']
    
    def _input_transformation(self, W, parameter):
        """Apply input transformation"""
        return W
    
    def _output_transformation(self, input_trans, parameter):
        """Apply output transformation"""
        return input_trans / len(self.parameters)

# Create instances for arguments
size = (3, 4, 5)
group = MockGroup()
new_size = [2, 3]
space = 'equivariant'

ARGS = {
    'size': size,
    'group': group,
    'new_size': new_size,
    'space': space
}
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
get_element_index Ran
ran on a synthesized input
ElisevanderPol/symmetrizer/symmetrizer/groups/ops.py
code served (permissive licence) · get_code("8e74655af7acc904")
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 int, ndim 0, dtype family int, 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 torch

torch.manual_seed(0)

# Create a small tensor to search for
el = torch.tensor([1.0, 2.0, 3.0, 4.0])

# Create a list of similar tensors
elements = [
    torch.tensor([1.0, 2.0, 3.0, 4.0]),  # This should match el
    torch.tensor([5.0, 6.0, 7.0, 8.0]),
    torch.tensor([9.0, 10.0, 11.0, 12.0]),
]

ARGS = {
    "el": el,
    "elements": elements,
}
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 int, ndim 0, dtype family int, 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
strip_ddp_state_dict Ran
ran; our contract draft was wrong, not the code
ElisevanderPol/mdp-homomorphic-networks/rlpyt/models/utils.py
code served (permissive licence) · get_code("1a929bb446f49571")
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 dict. 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 a sample state_dict with DDP-prefixed keys
state_dict = {
    'module.layer1.weight': torch.randn(4, 3),
    'module.layer1.bias': torch.randn(4),
    'module.layer2.weight': torch.randn(5, 4),
    'layer3.weight': torch.randn(2, 5),  # Already without 'module.' prefix
    'layer3.bias': torch.randn(2),
}

ARGS = {
    'state_dict': state_dict
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
symmetrize Ran
ran on a synthesized input
ElisevanderPol/symmetrizer/symmetrizer/ops/ops.py
code served (permissive licence) · get_code("772c37cc22f9d339")
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 (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)

# Create a small stand-in group class that implements the required interface
class SymmetryGroup:
    def __init__(self):
        self.parameters = [0, 1]  # Two parameters for the group
    
    def _input_transformation(self, W, parameter):
        """Transform input based on parameter"""
        # Simple transformation: scale by (parameter + 1)
        return W * (parameter + 1)
    
    def _output_transformation(self, transformed_input, parameter):
        """Transform output based on parameter"""
        # Simple transformation: scale by (parameter + 1)
        return transformed_input * (parameter + 1)

# Create a small weight matrix
W = torch.randn(4, 3)

# Create a group instance
group = SymmetryGroup()

ARGS = {
    "W": W,
    "group": group
}
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 (4, 3) — shape (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
symmetrize_invariant_out Ran
ran on a synthesized input
ElisevanderPol/symmetrizer/symmetrizer/ops/ops.py
code served (permissive licence) · get_code("c2487dde239bbdd1")
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 (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)

# Create a small stand-in class for the group parameter
class MockGroup:
    def __init__(self):
        self.parameters = [1, 2]  # Simple list of parameters
    
    def _output_transformation(self, W, parameter):
        """
        Mock transformation that scales W by the parameter value.
        This satisfies the function's requirement to call group._output_transformation(W, parameter)
        """
        return W * parameter

# Create a small weight matrix (4x3 to have distinct dimensions)
W = torch.randn(4, 3)

# Create the group object
group = MockGroup()

ARGS = {
    "W": W,
    "group": group
}
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 (4, 3) — shape (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

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 5 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
conv2d_output_shape 5 other papers
2 dated earlier, 3 later
1909.01500 (2019-09)
rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch
1910.13439 (2019-10)
Learning to Manipulate Deformable Objects without Demonstrations
2009.08319 (2020-09)
Decoupling Representation Learning from Reinforcement Learning
and 2 more not named here
strip_ddp_state_dict 5 other papers
2 dated earlier, 3 later
1909.01500 (2019-09)
rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch
1910.13439 (2019-10)
Learning to Manipulate Deformable Objects without Demonstrations
2009.08319 (2020-09)
Decoupling Representation Learning from Reinforcement Learning
and 2 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

This paper introduces MDP homomorphic networks for deep reinforcement learning. MDP homomorphic networks are neural networks that are equivariant under symmetries in the joint state-action space of an MDP. Current approaches to deep reinforcement learning do not usually exploit knowledge about such structure. By building this prior knowledge into policy and value networks using an equivariance constraint, we can reduce the size of the solution space. We specifically focus on group-structured symmetries (invertible transformations). Additionally, we introduce an easy method for constructing equivariant network layers numerically, so the system designer need not solve the constraints by hand, as is typically done. We construct MDP homomorphic MLPs and CNNs that are equivariant under either a group of reflections or rotations. We show that such networks converge faster than unstructured baselines on CartPole, a grid world and Pong.

For agents

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

get_harvested_code_for_paper("2006.16908")
get_code_for_paper("2006.16908")
have("2006.16908")

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

curl https://syntology.ai/api/ran/2006.16908.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.16908.svg)](https://syntology.ai/paper/2006.16908)

Connect an agent — have() is free.