SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2212.00124 · NeurIPS · 2023

One Risk to Rule Them All: A Risk-Sensitive Perspective on Model-Based Offline Reinforcement Learning

Marc Rigter, Bruno Lacerda, Nick Hawes

arXiv · PDF · Open in the Atlas

Code that ran

14 of 23 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 6 failures of our instrument; 9 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, 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/2212.00124.json · badge: /api/ran/2212.00124.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
marc-rigter/1R2R canonical (classifier label)
classifier basis: plan_row
5 ran · 3 no recorded run
nuria95/O-RAAC community
Papers with Code archive
6 ran · 2 no recorded run
copy not recorded community
not in the paper's declared repositories
2 ran · 0 no recorded run
microsoft/lightATAC community
Papers with Code archive
1 ran · 4 no recorded run
FunctionStatusWhere it lives
td_target Ran
ran, honoured the contract we drafted
marc-rigter/1R2R/_1R2R/algorithms/_1R2R.py
code served (permissive licence) · get_code("31deb9c7be58e00c")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T23:10:55Z). 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)

# Construct arguments for td_target(reward, discount, next_value)
# The function computes: reward + discount * next_value
# All parameters are numeric values or tensors

reward = torch.tensor([[1.0, 2.0, 3.0, 4.0],
                       [5.0, 6.0, 7.0, 8.0]])  # shape (2, 4)

discount = 0.99  # scalar discount factor

next_value = torch.tensor([[0.5, 0.6, 0.7, 0.8],
                           [0.9, 1.0, 1.1, 1.2]])  # shape (2, 4)

ARGS = {
    "reward": reward,
    "discount": discount,
    "next_value": next_value
}
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 Tensor, 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
CPW Ran
ran on a synthesized input · constructed an object; no result was computed
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("d354e29969a8eecf")
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 an instance of its own class, CPW: the run constructed the object and did not call it, so no result was computed. 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
from torch.distributions import uniform

torch.manual_seed(0)

ARGS = {
    "eta": 0.71
}
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 CPW — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CPW 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
Observation Ran
ran on a synthesized input · constructed an object; no result was computed
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("0a92b095c96f1001")
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 an instance of its own class, Observation: the run constructed the object and did not call it, so no result was computed. 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
from collections import namedtuple

class Observation(namedtuple('Observation', ('state', 'action', 'reward', 'next_state', 'done'))):
    """Observation datatype."""

    def __eq__(self, other):
        """Check if two observations are equal."""
        is_equal = np.allclose(self.state, other.state)
        is_equal &= np.allclose(self.action, other.action)
        is_equal &= np.allclose(self.reward, other.reward)
        is_equal &= np.allclose(self.next_state, other.next_state)
        is_equal &= self.done == other.done
        return is_equal

    def __ne__(self, other):
        """Check if two observations are not equal."""
        return not self == other


ARGS = {
    'state': np.array([0.1, 0.2, 0.3, 0.4]),
    'action': np.array([1.0, 2.0]),
    'reward': np.array(5.0),
    'next_state': np.array([0.2, 0.3, 0.4, 0.5]),
    'done': False
}
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 Observation — the run constructed an instance of its own class and did not call it; no shape 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
Power Ran
ran on a synthesized input · constructed an object; no result was computed
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("d0dc0545525ce106")
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 an instance of its own class, Power: the run constructed the object and did not call it, so no result was computed. 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
from torch.distributions import uniform
import torch

torch.manual_seed(0)

ARGS = {
    "eta": -2
}
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 Power — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Power 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
Wang_distortion Ran
ran on a synthesized input · constructed an object; no result was computed
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("f0d06853888e809c")
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 an instance of its own class, Wang_distortion: the run constructed the object and did not call it, so no result was computed. 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
from torch.distributions import uniform
from torch.distributions.normal import Normal

torch.manual_seed(0)

ARGS = {
    "eta": -0.75
}
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 Wang_distortion — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Wang_distortion 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
compute_cvar Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("0665de8d83eef80b")
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 float, shape (2,) 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 warnings

torch.manual_seed(0)

# Create sample data: batch_size=2, N=4 (number of samples per batch)
data = torch.randn(2, 4)

# alpha parameter: controls the threshold for CVaR computation
alpha = 0.3

ARGS = {
    "data": data,
    "alpha": alpha,
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) 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
format_samples_for_training Ran
ran; our contract draft was wrong, not the code
marc-rigter/1R2R/_1R2R/models/constructor.py
code served (permissive licence) · get_code("1bb5583dba32f555")
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 sample data with distinct dimensions
batch_size = 2
obs_dim = 4
action_dim = 3
reward_dim = 1

# Create observations: shape (batch_size, obs_dim)
observations = np.random.randn(batch_size, obs_dim).astype(np.float32)

# Create actions: shape (batch_size, action_dim)
actions = np.random.randn(batch_size, action_dim).astype(np.float32)

# Create next_observations: shape (batch_size, obs_dim)
next_observations = np.random.randn(batch_size, obs_dim).astype(np.float32)

# Create rewards: shape (batch_size, reward_dim)
rewards = np.random.randn(batch_size, reward_dim).astype(np.float32)

# Construct the samples dict
samples = {
    'observations': observations,
    'actions': actions,
    'next_observations': next_observations,
    'rewards': rewards
}

ARGS = {
    'samples': samples
}
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
get_required_argument Ran
ran; our contract draft was wrong, not the code
marc-rigter/1R2R/_1R2R/models/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from quanvuong/handful-of-trials-pytorch, is pointer only (licence: NONE) · get_code("4eb5226056a1f70f")
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 quanvuong/handful-of-trials-pytorch, 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.

l2_projection Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from chinganc/lightatac
pointer only · get_code("36c97c87a5d855c4")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for 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.

mass_center Ran
ran on a synthesized input
marc-rigter/1R2R/_1R2R/env/humanoid.py
code served (permissive licence) · get_code("46bd205700c42f20")
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 float64, 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

class MockModel:
    def __init__(self):
        # body_mass should be a 1D array of masses for each body
        # Using 4 bodies with distinct masses
        self.body_mass = np.array([1.0, 2.0, 3.0, 4.0])

class MockSimData:
    def __init__(self):
        # xipos should be shape (n_bodies, 3) for 3D positions
        # 4 bodies, 3 spatial dimensions
        self.xipos = np.array([
            [1.0, 2.0, 3.0],
            [4.0, 5.0, 6.0],
            [7.0, 8.0, 9.0],
            [10.0, 11.0, 12.0]
        ])

class MockSim:
    def __init__(self):
        self.data = MockSimData()

model = MockModel()
sim = MockSim()

ARGS = {
    "model": model,
    "sim": sim
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
normalized_sum Ran
ran, honoured the contract we drafted
this paper's copy was not recorded; identical code first harvested from microsoft/atac
pointer only · get_code("2a76fc3a7969ba54")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for 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.

quantile_huber_loss Ran
ran; our contract draft was wrong, not the code
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("24475939b27492ce")
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 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 torch.nn.functional as F

torch.manual_seed(0)

batch_size = 2
num_quantiles = 4

# T_theta: Target quantiles of size [batch_size x num_quantiles]
T_theta = torch.randn(batch_size, num_quantiles)

# Theta: Current quantiles of size [batch_size x num_quantiles]
Theta = torch.randn(batch_size, num_quantiles)

# tau_quantiles: Quantile levels of size [1 x num_quantiles]
tau_quantiles = torch.linspace(0.1, 0.9, num_quantiles).unsqueeze(0)

# k parameter (default value)
k = 1

ARGS = {
    "T_theta": T_theta,
    "Theta": Theta,
    "tau_quantiles": tau_quantiles,
    "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 Tensor, 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
restore_pool Ran
ran on a synthesized input
marc-rigter/1R2R/_1R2R/off_policy/loader.py
code served (permissive licence) · get_code("721b133d3e718b81")
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 2, 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 os
import pickle
import numpy as np
import tempfile

# Create a minimal ReplayPool class that the function will use
class ReplayPool:
    def __init__(self):
        self.size = 0
        self.data = {}
    
    def add_samples(self, data):
        # Simulate adding samples to the pool
        self.size += data['observations'].shape[0]
        self.data = data

# Create a minimal IllustrativeExample class for the gen_dataset function
class IllustrativeExample:
    def generate_dataset(self):
        num_samples = 100
        obs_dim = 4
        action_dim = 2
        return {
            'observations': np.random.randn(num_samples, obs_dim).astype(np.float32),
            'next_observations': np.random.randn(num_samples, obs_dim).astype(np.float32),
            'actions': np.random.randn(num_samples, action_dim).astype(np.float32),
            'rewards': np.random.randn(num_samples, 1).astype(np.float32),
            'terminals': np.zeros((num_samples, 1), dtype=np.float32)
        }

# Create a temporary directory with a mock dataset file
temp_dir = tempfile.mkdtemp()
mock_dataset = {
    'observations': np.random.randn(50, 4).astype(np.float32),
    'next_observations': np.random.randn(50, 4).astype(np.float32),
    'actions': np.random.randn(50, 2).astype(np.float32),
    'rewards': np.random.randn(50, 1).astype(np.float32),
    'terminals': np.zeros((50, 1), dtype=np.float32)
}
dataset_path = os.path.join(temp_dir, 'test_dataset.pkl')
with open(dataset_path, 'wb') as f:
    pickle.dump(mock_dataset, f)

# Construct the arguments
replay_pool = ReplayPool()
experiment_root = 'datasets/test_dataset'
save_path = None
normalize_states = True
normalize_rewards = False
dataset_dir = temp_dir

ARGS = {
    'replay_pool': replay_pool,
    'experiment_root': experiment_root,
    'save_path': save_path,
    'normalize_states': normalize_states,
    'normalize_rewards': normalize_rewards,
    'dataset_dir': dataset_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
d4rl, gym, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, 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
split Ran
ran; our contract draft was wrong, not the code
microsoft/lightATAC/lightATAC/atac.py
code served (permissive licence) · get_code("fd64627ef6698537")
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 generator. 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.distributions as dist

torch.manual_seed(0)

# Create a simple Normal distribution with batch shape (2,)
loc = torch.randn(2, 4)
scale = torch.ones(2, 4)
normal_dist = dist.Normal(loc, scale)

# Test with integer split size
split_size_or_sections_int = 2

ARGS = {
    "tensor_or_distribution": normal_dist,
    "split_size_or_sections": split_size_or_sections_int,
}
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 generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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
ATAC Not yet run
unverified (harvested, no recorded run)
microsoft/lightATAC/lightATAC/atac.py
code served (permissive licence) · get_code("385289a2f103a41b")
How it ran

no invocation recorded.

AbstractAgent Not yet run
unverified (harvested, no recorded run)
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("c5bf09ce02ff5964")
How it ran

no invocation recorded.

ORAAC Not yet run
unverified (harvested, no recorded run)
nuria95/O-RAAC/oraaclib/agent/ORAAC.py
code served (permissive licence) · get_code("7d39619c9ae882e2")
How it ran

no invocation recorded.

compute_batched Not yet run
unverified (harvested, no recorded run)
microsoft/lightATAC/lightATAC/atac.py
code served (permissive licence) · get_code("95422abceacb6468")
How it ran

no invocation recorded.

gen_dataset Not yet run
unverified (harvested, no recorded run)
marc-rigter/1R2R/_1R2R/off_policy/loader.py
code served (permissive licence) · get_code("b69606fbe4ae1c6d")
How it ran

no invocation recorded.

restore_pool_d4rl Not yet run
unverified (harvested, no recorded run)
marc-rigter/1R2R/_1R2R/off_policy/loader.py
code served (permissive licence) · get_code("4a8a13bf2658940e")
How it ran

no invocation recorded.

slice_dist Not yet run
unverified (harvested, no recorded run)
microsoft/lightATAC/lightATAC/atac.py
code served (permissive licence) · get_code("d5ae5ee54ffb75d0")
How it ran

no invocation recorded.

spectral_norm Not yet run
unverified (harvested, no recorded run)
marc-rigter/1R2R/_1R2R/models/utils.py
code served (permissive licence) · get_code("633dcc6fe7c3ecc3")
How it ran

no invocation recorded.

split_dist Not yet run
unverified (harvested, no recorded run)
microsoft/lightATAC/lightATAC/atac.py
code served (permissive licence) · get_code("c9923434b7a879c9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 2 later) and 17 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
td_target 8 other papers
8 dated earlier
1703.03400 (2017-03)
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
1812.05905 (2018-12)
Soft Actor-Critic Algorithms and Applications
1906.08253 (2019-06)
When to Trust Your Model: Model-Based Policy Optimization
and 5 more not named here
format_samples_for_training 7 other papers
6 dated earlier, 1 later · 1 only as an assembly brick
2005.13239 (2020-05)
MOPO: Model-based Offline Policy Optimization
2006.13916 (2020-06)
Off-Dynamics Reinforcement Learning: Training for Transfer with Domain Classifiers
2010.09546 (2020-10)
Model-based Policy Optimization with Unsupervised Model Adaptation
and 4 more not named here
get_required_argument 6 other papers
6 dated earlier
1805.12114 (2018-05)
Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
2006.13916 (2020-06)
Off-Dynamics Reinforcement Learning: Training for Transfer with Domain Classifiers
2007.01995 (2020-07)
Bidirectional Model-based Policy Optimization
and 3 more not named here
mass_center 4 other papers
4 dated earlier
2006.13916 (2020-06)
Off-Dynamics Reinforcement Learning: Training for Transfer with Domain Classifiers
2103.13842 (2021-03)
Model Predictive Actor-Critic: Accelerating Robot Skill Acquisition with Deep Reinforcement Learning
2104.06922 (2021-04)
Safe Continuous Control with Constrained Model-Based Policy Optimization
and 1 more not named here
l2_projection
brick of this paper's assembly: helper
2 other papers
1 dated earlier, 1 later
2202.02446 (2022-02)
Adversarially Trained Actor Critic for Offline Reinforcement Learning
2301.12714 (2023-01)
Importance Weighted Actor-Critic for Optimal Conservative Offline Reinforcement Learning
normalized_sum
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2202.02446 (2022-02)
Adversarially Trained Actor Critic for Offline Reinforcement Learning

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

Offline reinforcement learning (RL) is suitable for safety-critical domains where online exploration is not feasible. In such domains, decision-making should take into consideration the risk of catastrophic outcomes. In other words, decision-making should be risk-averse. An additional challenge of offline RL is avoiding distributional shift, i.e. ensuring that state-action pairs visited by the policy remain near those in the dataset. Previous offline RL algorithms that consider risk combine offline RL techniques (to avoid distributional shift), with risk-sensitive RL algorithms (to achieve risk-aversion). In this work, we propose risk-aversion as a mechanism to jointly address both of these issues. We propose a model-based approach, and use an ensemble of models to estimate epistemic uncertainty, in addition to aleatoric uncertainty. We train a policy that is risk-averse, and avoids high uncertainty actions. Risk-aversion to epistemic uncertainty prevents distributional shift, as areas not covered by the dataset have high epistemic uncertainty. Risk-aversion to aleatoric uncertainty discourages actions that are risky due to environment stochasticity. Thus, by considering epistemic uncertainty via a model ensemble and introducing risk-aversion, our algorithm (1R2R) avoids distributional shift in addition to achieving risk-aversion to aleatoric risk. Our experiments show that 1R2R achieves strong performance on deterministic benchmarks, and outperforms existing approaches for risk-sensitive objectives in stochastic domains.

For agents

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

get_harvested_code_for_paper("2212.00124")
get_code_for_paper("2212.00124")
have("2212.00124")

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

curl https://syntology.ai/api/ran/2212.00124.json

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

[![Syntology run record](https://syntology.ai/api/ran/2212.00124.svg)](https://syntology.ai/paper/2212.00124)

Connect an agent — have() is free.