SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1802.09477 · 2018

Addressing Function Approximation Error in Actor-Critic Methods

David Meger, Scott Fujimoto, Herke Van Hoof

arXiv · PDF · Open in the Atlas

Code that ran

26 of 36 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, 2 violated it and 17 have no contract verdict; 6 failures of our instrument; 10 no recorded run), measured 2026-11-20; 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/1802.09477.json · badge: /api/ran/1802.09477.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
robintyh1/icml2021-pengqlambda community (classifier label)
classifier basis not recorded on the link
7 ran · 4 no recorded run
DanielTakeshi/DCUR community (classifier label)
classifier basis not recorded on the link
6 ran · 3 no recorded run
jaem-seo/AI_tokamak_control community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
ollenilsson19/MAP-Elites-GAPG community (classifier label)
classifier basis not recorded on the link
3 ran · 2 no recorded run
yydsok/oparl community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
markub3327/rl-agent community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
VasaKiDD/TD3-deep-rl-research extension (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
yifan12wu/td3-jax community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
actv Ran
ran on a synthesized input
jaem-seo/AI_tokamak_control/common/model_structure.py
code served (permissive licence) · get_code("b798db60af40d446")
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 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 numpy as np

# Create a small test input array
x = np.array([[-2.0, -1.0, 0.5, 2.0],
              [1.5, -0.5, 0.0, 3.0]], dtype=np.float32)

# Test each method
method = 'relu'

ARGS = {
    'x': x,
    'method': method
}
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 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
all_bools Ran
ran, violated the contract we drafted
robintyh1/icml2021-pengqlambda/spinup/utils/run_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("2103e65a90ce7a04")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

colorize Ran
ran; our contract draft was wrong, not the code
robintyh1/icml2021-pengqlambda/spinup/utils/logx.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("7e36f1c357c4618d")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

convert_json Ran
ran; our contract draft was wrong, not the code
robintyh1/icml2021-pengqlambda/spinup/utils/serialization_utils.py
pointer only (licence: MIT) · get_code("fbc505a5dac82646")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

cosine_schedule Ran
ran on a synthesized input
markub3327/rl-agent/rl_toolkit/networks/callbacks/lr.py
code served (permissive licence) · get_code("38aa9c2f06df49b7")
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 function. 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 tensorflow as tf

ARGS = {
    "base_lr": 0.001,
    "total_steps": 100,
    "warmup_steps": 10
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
density_scatter Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/inspect_policy.py
pointer only (licence: MIT) · get_code("e0f9ede3192984c6")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

evaluate_policy Ran
ran, honoured the contract we drafted
VasaKiDD/TD3-deep-rl-research/train/main_faimemory.py
code served (permissive licence) · get_code("b0a2d4b21b2d48e6")
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) and returned float, 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
import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal policy class that has select_action method
class MinimalPolicy(nn.Module):
    def __init__(self, obs_dim=4, action_dim=2):
        super().__init__()
        self.linear = nn.Linear(obs_dim, action_dim)
    
    def select_action(self, obs):
        # obs is a numpy array, convert to tensor
        obs_tensor = torch.FloatTensor(obs).unsqueeze(0)
        with torch.no_grad():
            action = self.linear(obs_tensor)
        # Return a scalar action (or array depending on action space)
        return action.squeeze(0).numpy()

# Create a minimal environment class
class MinimalEnv:
    def __init__(self, obs_dim=4, action_dim=2):
        self.obs_dim = obs_dim
        self.action_dim = action_dim
        self.step_count = 0
        self.max_steps = 5
    
    def reset(self):
        self.step_count = 0
        return np.random.randn(self.obs_dim).astype(np.float32)
    
    def step(self, action):
        self.step_count += 1
        obs = np.random.randn(self.obs_dim).astype(np.float32)
        reward = float(np.random.randn())
        done = self.step_count >= self.max_steps
        info = {}
        return obs, reward, done, info

# Create instances
policy = MinimalPolicy(obs_dim=4, action_dim=2)
env = MinimalEnv(obs_dim=4, action_dim=2)

# Make env globally accessible for the function
import sys
current_module = sys.modules[__name__]
current_module.env = env

ARGS = {
    'policy': policy,
    'eval_episodes': 10
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float — no output shape is stated: observe_signature.py could not observe it (fixture:KeyError)
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
f2i Ran
ran on a synthesized input
jaem-seo/AI_tokamak_control/ai_control_v0.py
code served (permissive licence) · get_code("ff2f8ee596f45c07")
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; no output shape is stated. It ran in syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1) at 2026-09-26T21:51:12Z. 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

decimals = np.log10(1000)

def f2i(f, decimals=decimals):
    return int(f * 10 ** decimals)

ARGS = {
    "f": 3.14159,
    "decimals": decimals
}
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 int, ndim 0, dtype family int — no output shape is stated: no later run recorded an output shape for this sample (observed_output and metamorphic_properties carry none; shape_contract and output_contract record ndim only)
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 syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1); 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
generate_grid_configs Ran
ran on a synthesized input
ollenilsson19/MAP-Elites-GAPG/configure_experiment/generate_configs.py
code served (permissive licence) · get_code("df403533c9480409")
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 list, ndim 1, 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.

from sklearn.model_selection import ParameterGrid

torch_imported = False
try:
    import torch
    torch.manual_seed(0)
    torch_imported = True
except ImportError:
    pass

# Analyze the function:
# - varables: a list of variable names (strings)
# - ranges: a list of ranges (each element corresponds to a variable in varables)
# The function zips them together into a param_grid dict and returns ParameterGrid results

varables = ["learning_rate", "batch_size", "num_layers"]
ranges = [
    [0.001, 0.01],
    [16, 32, 64],
    [2, 3]
]

ARGS = {
    "varables": varables,
    "ranges": ranges
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
sklearn, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list 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_all_datasets Ran
ran; our contract draft was wrong, not the code
robintyh1/icml2021-pengqlambda/spinup/utils/plot.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("62314e3b8ca6edfc")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_base_conf Ran
ran on a synthesized input
ollenilsson19/MAP-Elites-GAPG/configure_experiment/generate_configs.py
code served (permissive licence) · get_code("87cf4755f4594737")
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 tempfile
import os

# Create a temporary config file with valid command-line style arguments
temp_dir = tempfile.mkdtemp()
config_file = os.path.join(temp_dir, "base_config.txt")

with open(config_file, 'w') as f:
    f.write("--learning_rate=0.001\n")
    f.write("--batch_size=32\n")
    f.write("--num_epochs=10\n")
    f.write("--use_dropout\n")
    f.write("--model_name=bert\n")

ARGS = {
    "base_config": config_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
get_datasets Ran
ran; our contract draft was wrong, not the code
robintyh1/icml2021-pengqlambda/spinup/utils/plot.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("e642bb98eca55069")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_revised_reward Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/offline_utils.py
pointer only (licence: MIT) · get_code("bd6be7742940fb51")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

i2f Ran
ran on a synthesized input
jaem-seo/AI_tokamak_control/ai_control_v0.py
code served (permissive licence) · get_code("5fa1289490ba5c13")
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 float, 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

decimals = np.log10(1000)

def i2f(i, decimals=decimals):
    return float(i / 10 ** decimals)

ARGS = {
    "i": 500,
    "decimals": decimals
}
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 float, 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
is_json_serializable Ran
ran, violated the contract we drafted
robintyh1/icml2021-pengqlambda/spinup/utils/serialization_utils.py
pointer only (licence: MIT) · get_code("9ca8b8e45f7fbebf")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

make_hashable Ran
ran; our contract draft was wrong, not the code
ollenilsson19/MAP-Elites-GAPG/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from adaptive-intelligent-robotics/moqd-csp, is pointer only (licence: NONE) · get_code("c17b6915065d3722")
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 adaptive-intelligent-robotics/moqd-csp, and GitHub detected no licence in that repository when it was checked. 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.

plot_fill Ran
ran on a synthesized input
yifan12wu/td3-jax/plot.py
code served (permissive licence) · get_code("986b521c8e695d46")
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 Line2D. 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 matplotlib.pyplot as plt

torch.manual_seed(0)

# Create simple x, y, and yerr arrays for plotting
x = np.linspace(0, 10, 5)
y = np.sin(x)
yerr = np.abs(np.random.randn(5) * 0.1)

# Create a figure and axis
fig, ax = plt.subplots()

ARGS = {
    'x': x,
    'y': y,
    'yerr': yerr,
    'color': 'blue',
    'alpha_fill': 0.2,
    'ax': ax,
    'label': 'test data',
    'ls': '-'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Line2D — a Line2D output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Line2D 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
r2_k Ran
ran on a synthesized input
jaem-seo/AI_tokamak_control/old/model_structure.py
code served (permissive licence) · get_code("f7d8b73059cc19fc")
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 float, 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)

# The function r2_k takes y_true and y_pred as parameters
# Based on the function signature, both should be array-like (tensors or numpy arrays)
# that can be compared for regression metrics

y_true = torch.randn(2, 4)
y_pred = torch.randn(2, 4)

ARGS = {
    "y_true": y_true,
    "y_pred": y_pred,
}
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 float, 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
sample_combo_batch Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/online_rl.py
pointer only (licence: MIT) · get_code("7a6fd6ea6db8460a")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

sample_std Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/load_policy.py
pointer only (licence: MIT) · get_code("1b7cc1754bd06854")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

should_we_end_early Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/offline_utils.py
pointer only (licence: MIT) · get_code("fb3b17a4bd91d20a")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

smooth Ran
ran on a synthesized input
DanielTakeshi/DCUR/spinup/teaching/inspect_actions.py
pointer only (licence: MIT) · get_code("c9672c271269cb9d")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

split_list Ran
ran on a synthesized input
yydsok/oparl/parell.py
code served (permissive licence) · get_code("1dcf318473b1f057")
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 list, ndim 2, 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

torch.manual_seed(0)

# Define arguments for split_list function
input_list = [1, 2, 3, 4, 5, 6, 7, 8, 9]
K = 3

ARGS = {
    "input_list": input_list,
    "K": K
}
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 list, ndim 2, dtype family int — a list 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
tensormax Ran
ran on a synthesized input
yydsok/oparl/TD3.py
code served (permissive licence) · get_code("052527002e0be3dc")
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 (1, 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)

# The function iterates over targetQ and compares tensors using torch.gt
# It expects targetQ to be an iterable of tensors
# torch.gt returns a boolean tensor, and [0] indexes into it
# The function compares tensors element-wise and selects the maximum

targetQ = [
    torch.tensor([[1.0, 2.0, 3.0, 4.0]]),
    torch.tensor([[2.0, 1.0, 4.0, 3.0]]),
]

ARGS = {
    "targetQ": targetQ
}
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 (1, 4) — shape (1, 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
tensormin Ran
ran on a synthesized input
yydsok/oparl/TD3.py
code served (permissive licence) · get_code("a8d4fda0f5e30ac7")
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; no output shape is stated. It ran in syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1) at 2026-09-26T21:51:12Z. 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

torch.manual_seed(0)

# The function expects targetQ to be a list/sequence of tensors
# It iterates through them and uses torch.lt for comparison
# Each tensor needs to support indexing with [0]
# Based on the logic, targetQ should be a list of 1D tensors (or tensors where [0] gives a scalar)

targetQ = [
    torch.tensor([5.0]),
    torch.tensor([3.0]),
    torch.tensor([7.0]),
    torch.tensor([2.0])
]

ARGS = {
    "targetQ": targetQ
}
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 1, dtype family float — no output shape is stated: no later run recorded an output shape for this sample (observed_output and metamorphic_properties carry none; shape_contract and output_contract record ndim only)
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 syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1); 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
valid_str Ran
ran; our contract draft was wrong, not the code
robintyh1/icml2021-pengqlambda/spinup/utils/run_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("c6bdd181afa1fdf7")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

add_to_archive Not yet run
unverified (harvested, no recorded run)
ollenilsson19/MAP-Elites-GAPG/utils.py
code served (permissive licence) · get_code("c005d4991ad325bd")
How it ran

no invocation recorded.

cvt Not yet run
unverified (harvested, no recorded run)
ollenilsson19/MAP-Elites-GAPG/utils.py
code served (permissive licence) · get_code("b97fcec4203f4736")
How it ran

no invocation recorded.

get_vars Not yet run
unverified (harvested, no recorded run)
robintyh1/icml2021-pengqlambda/spinup/algos/tf1/td3_peng/core.py
pointer only (licence: MIT) · get_code("a2580d7c4a763348")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

load_custom_model Not yet run
unverified (harvested, no recorded run)
jaem-seo/AI_tokamak_control/common/model_structure.py
code served (permissive licence) · get_code("2d2a11f97ea11eec")
How it ran

no invocation recorded.

load_policy_and_env Not yet run
unverified (harvested, no recorded run)
DanielTakeshi/DCUR/spinup/teaching/load_policy.py
pointer only (licence: MIT) · get_code("a921320c1a9ef680")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

load_policy_model Not yet run
unverified (harvested, no recorded run)
DanielTakeshi/DCUR/spinup/teaching/inspect_actions.py
pointer only (licence: MIT) · get_code("de91c09e463e0669")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

load_pytorch_policy Not yet run
unverified (harvested, no recorded run)
DanielTakeshi/DCUR/spinup/teaching/load_policy.py
pointer only (licence: MIT) · get_code("e50af51bd256e662")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, DanielTakeshi/DCUR, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

mlp Not yet run
unverified (harvested, no recorded run)
robintyh1/icml2021-pengqlambda/spinup/algos/tf1/td3_peng/core.py
pointer only (licence: MIT) · get_code("e68b07bcdded00fa")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

placeholder Not yet run
unverified (harvested, no recorded run)
robintyh1/icml2021-pengqlambda/spinup/algos/tf1/td3_peng/core.py
pointer only (licence: MIT) · get_code("17b03c462ce92576")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

restore_tf_graph Not yet run
unverified (harvested, no recorded run)
robintyh1/icml2021-pengqlambda/spinup/utils/logx.py
pointer only (licence: MIT); the stored copy get_code serves, from milanganai/milanganai.github.io, is pointer only (licence: NONE) · get_code("8a70bbd1276e274c")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, robintyh1/icml2021-pengqlambda, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

Also present in other papers' code

Of 36 samples lifted from this paper's linked repositories, 36 carry a function identity and 0 do not. Of the 36 distinct functions with an identity, 13 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 13 later) and 23 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
colorize 17 other papers
17 later
1909.12255 (2019-09)
Harnessing Structures for Value-Based Planning and Reinforcement Learning
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
and 14 more not named here
is_json_serializable 9 other papers
9 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
2205.14691 (2022-05)
On the Robustness of Safe Reinforcement Learning under Observational Perturbations
and 6 more not named here
convert_json 7 other papers
7 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
2205.14691 (2022-05)
On the Robustness of Safe Reinforcement Learning under Observational Perturbations
and 4 more not named here
mlp 5 other papers
1 dated earlier, 4 later
1801.01290 (2018-01)
Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
1812.02690 (2018-12)
Provably Efficient Maximum Entropy Exploration
1910.07224 (2019-10)
Teacher algorithms for curriculum learning of Deep RL in continuously parameterized environments
and 2 more not named here
get_all_datasets 4 other papers
4 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
2310.13230 (2023-10)
Absolute Policy Optimization
and 1 more not named here
get_vars 4 other papers
4 later
1812.02690 (2018-12)
Provably Efficient Maximum Entropy Exploration
1910.07224 (2019-10)
Teacher algorithms for curriculum learning of Deep RL in continuously parameterized environments
2302.00808 (2023-02)
ACPO: A Policy Optimization Algorithm for Average MDPs with Constraints
and 1 more not named here
placeholder 4 other papers
4 later
1812.02690 (2018-12)
Provably Efficient Maximum Entropy Exploration
1910.07224 (2019-10)
Teacher algorithms for curriculum learning of Deep RL in continuously parameterized environments
2302.00808 (2023-02)
ACPO: A Policy Optimization Algorithm for Average MDPs with Constraints
and 1 more not named here
restore_tf_graph 4 other papers
4 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2210.07573 (2022-10)
Model-based Safe Deep Reinforcement Learning via a Constrained Proximal Policy Optimization Algorithm
2605.09638 (2026-05)
Plan2Cleanse: Test-Time Backdoor Defense via Monte-Carlo Planning in Deep Reinforcement Learning
and 1 more not named here
valid_str 4 other papers
4 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.04619 (2021-06)
Self-Supervised Learning with Data Augmentations Provably Isolates Content from Style
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
and 1 more not named here
all_bools 3 other papers
3 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2106.09110 (2021-06)
Safe Reinforcement Learning Using Advantage-Based Intervention
2605.11975 (2026-05)
Stochastic Minimum-Cost Reach-Avoid Reinforcement Learning
get_datasets 2 other papers
2 later
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
2605.11975 (2026-05)
Stochastic Minimum-Cost Reach-Avoid Reinforcement Learning
cosine_schedule 1 other paper
1 later
2005.05719 (2020-05)
Smooth Exploration for Robotic Reinforcement Learning
make_hashable 1 other paper
1 later
2403.17164 (2024-03)
Multi-Objective Quality-Diversity for Crystal Structure Prediction

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 2026-11-20.

Repositories linked to this paper

More are linked than shown here.

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

Abstract

In value-based reinforcement learning methods such as deep Q-learning, function approximation errors are known to lead to overestimated value estimates and suboptimal policies. We show that this problem persists in an actor-critic setting and propose novel mechanisms to minimize its effects on both the actor and the critic. Our algorithm builds on Double Q-learning, by taking the minimum value between a pair of critics to limit overestimation. We draw the connection between target networks and overestimation bias, and suggest delaying policy updates to reduce per-update error and further improve performance. We evaluate our method on the suite of OpenAI gym tasks, outperforming the state of the art in every environment tested.

For agents

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

get_harvested_code_for_paper("1802.09477")
get_code_for_paper("1802.09477")
have("1802.09477")

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

curl https://syntology.ai/api/ran/1802.09477.json

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

[![Syntology run record](https://syntology.ai/api/ran/1802.09477.svg)](https://syntology.ai/paper/1802.09477)

Connect an agent — have() is free.