SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2207.08894 · 2022

A Deep Reinforcement Learning Approach for Finding Non-Exploitable Strategies in Two-Player Atari Games

arXiv · PDF · Open in the Atlas

Code that ran

5 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 5 have no contract verdict; 0 failures of our instrument; 5 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/2207.08894.json · badge: /api/ran/2207.08894.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
quantumiracle/mars canonical (classifier label)
classifier basis not recorded on the link
5 ran · 5 no recorded run
FunctionStatusWhere it lives
NashEquilibriumCVXPYSolver Ran
ran on a synthesized input
quantumiracle/mars/mars/equilibrium_solver/eq_CVXPYsolver.py
code served (permissive licence) · get_code("81e21540e0757e7b")
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

# Create a small payoff matrix for a 2-player game
# Rows represent player 1's strategies, columns represent player 2's strategies
A = np.array([
    [3.0, 1.0, 2.0],
    [2.0, 4.0, 1.0],
    [1.0, 2.0, 3.0]
], dtype=np.float64)

ARGS = {
    "A": A,
    "verbose": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cvxpy, 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
arrayWithoutElement Ran
ran on a synthesized input
quantumiracle/mars/mars/equilibrium_solver/gamegenerator.py
code served (permissive licence) · get_code("50b1e533c45c2ec3")
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 int, shape (8,) 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

# Construct arguments for arrayWithoutElement
# The function takes:
# - A: a numpy array
# - i: an index to exclude

# Create a small numpy array
A = np.array([[1, 2, 3, 4],
              [5, 6, 7, 8],
              [9, 10, 11, 12]])

# Index to exclude (row 1)
i = 1

ARGS = {
    "A": A,
    "i": i
}
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 int, shape (8,) — shape (8,) 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
cartesian Ran
ran on a synthesized input
quantumiracle/mars/mars/equilibrium_solver/gamegenerator.py
code served (permissive licence) · get_code("1e9c1d4bf7ca8df8")
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 2, dtype family int, shape (24, 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 numpy as np

# Create sample arrays for the cartesian product function
array1 = np.array([1, 2, 3])
array2 = np.array([4, 5])
array3 = np.array([6, 7, 8, 9])

arrays = [array1, array2, array3]

ARGS = {
    "arrays": arrays,
    "out": None
}
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 2, dtype family int, shape (24, 3) — shape (24, 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
get_payoff_vector Ran
ran on a synthesized input
quantumiracle/mars/mars/equilibrium_solver/eq_MWUsolver.py
code served (permissive licence) · get_code("f03da9b695af9788")
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 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 torch
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# - payoff_matrix: used with @ operator (matrix multiplication)
# - opponent_policy: used with @ operator (matrix multiplication)
# The operation is: opponent_policy @ payoff_matrix
# So opponent_policy should be 1D or 2D, and payoff_matrix should be 2D
# Result is payoff which is 1D or 2D

# opponent_policy: shape (num_actions,) or (batch, num_actions)
# payoff_matrix: shape (num_actions, num_outcomes)
# payoff result: shape (num_outcomes,) or (batch, num_outcomes)

opponent_policy = torch.tensor([0.3, 0.7], dtype=torch.float32)  # shape (2,)
payoff_matrix = torch.tensor([[1.0, 2.0, 3.0, 4.0],
                               [5.0, 6.0, 7.0, 8.0]], dtype=torch.float32)  # shape (2, 4)

ARGS = {
    "payoff_matrix": payoff_matrix,
    "opponent_policy": opponent_policy,
}
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 Tensor, 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
parseGame Ran
ran on a synthesized input
quantumiracle/mars/mars/equilibrium_solver/gamegenerator.py
code served (permissive licence) · get_code("90d75e71673e0ce9")
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

# Construct a small game matrix A
# The function parseGame expects A to be an array where:
# - shape[0] is the number of players
# - shape[1:] are the strategy dimensions for each player
# For a 2-player game with player 1 having 3 strategies and player 2 having 4 strategies:
A = np.random.rand(2, 3, 4)

ARGS = {
    "A": 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 — 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
CoarseCorrelatedEquilibriumLPSolver Not yet run
unverified (harvested, no recorded run)
quantumiracle/mars/mars/equilibrium_solver/eq_LPsolver.py
code served (permissive licence) · get_code("250b4ac1fe45568c")
How it ran

no invocation recorded.

NashEquilibriumLPSolver Not yet run
unverified (harvested, no recorded run)
quantumiracle/mars/mars/equilibrium_solver/eq_LPsolver.py
code served (permissive licence) · get_code("a3d4bff369d05675")
How it ran

no invocation recorded.

NashEquilibriumMWUSolver Not yet run
unverified (harvested, no recorded run)
quantumiracle/mars/mars/equilibrium_solver/eq_MWUsolver.py
code served (permissive licence) · get_code("7ce310a0a27122d0")
How it ran

no invocation recorded.

NashEquilibriumParallelMWUSolver Not yet run
unverified (harvested, no recorded run)
quantumiracle/mars/mars/equilibrium_solver/eq_MWUsolver.py
code served (permissive licence) · get_code("e5fe2f8ec0681121")
How it ran

no invocation recorded.

make_env Not yet run
unverified (harvested, no recorded run)
quantumiracle/mars/mars/env/import_env.py
code served (permissive licence) · get_code("7bd11d0025602c10")
How it ran

no invocation recorded.

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

This paper proposes new, end-to-end deep reinforcement learning algorithms for learning two-player zero-sum Markov games. Different from prior efforts on training agents to beat a fixed set of opponents, our objective is to find the Nash equilibrium policies that are free from exploitation by even the adversarial opponents. We propose (a) Nash-DQN algorithm, which integrates the deep learning techniques from single DQN into the classic Nash Q-learning algorithm for solving tabular Markov games; (b) Nash-DQN-Exploiter algorithm, which additionally adopts an exploiter to guide the exploration of the main agent. We conduct experimental evaluation on tabular examples as well as various two-player Atari games. Our empirical results demonstrate that (i) the policies found by many existing methods including Neural Fictitious Self Play and Policy Space Response Oracle can be prone to exploitation by adversarial opponents; (ii) the output policies of our algorithms are robust to exploitation, and thus outperform existing methods.

For agents

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

get_harvested_code_for_paper("2207.08894")
get_code_for_paper("2207.08894")
have("2207.08894")

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

curl https://syntology.ai/api/ran/2207.08894.json

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

[![Syntology run record](https://syntology.ai/api/ran/2207.08894.svg)](https://syntology.ai/paper/2207.08894)

Connect an agent — have() is free.