SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.14154 · NeurIPS · 2020

Strictly Batch Imitation Learning by Energy-based Distribution Matching

Mihaela Van Der Schaar, Daniel Jarrett, Ioana Bica

arXiv · PDF · Open in the Atlas

Code that ran

16 of 17 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (5 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 13 have no contract verdict; 3 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2006.14154.json · badge: /api/ran/2006.14154.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
dtak/batch-apprenticeship-learning canonical (classifier label)
classifier basis: census
9 ran · 1 no recorded run
wgrathwohl/JEM community
Papers with Code archive
7 ran · 0 no recorded run
FunctionStatusWhere it lives
ActNorm Ran
ran on a synthesized input · constructed an object; no result was computed
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("c22f2d08409192a0")
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, ActNorm: 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 as t
import torch.nn as nn
import torch.nn.functional as tnnF
import torch.distributions as tdist
import torch

torch.manual_seed(0)

ARGS = {
    'in_channel': 4,
    'logdet': True
}
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 ActNorm — 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 ActNorm 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
F Ran
ran on a synthesized input · constructed an object; no result was computed
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("4985cbdb3ad39e8c")
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, F: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'depth': 28,
    'width': 2,
    'norm': None,
    'dropout_rate': 0.0,
    'n_classes': 10
}
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 F — 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 F 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
Identity Ran
ran on a synthesized input · constructed an object; no result was computed
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("fdbb7a963e661fb1")
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, Identity: 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 as t
import torch.nn as nn
import torch.nn.functional as tnnF
import torch.distributions as tdist

t.manual_seed(0)

class Identity(nn.Module):

    def __init__(self, *args, **kwargs):
        super().__init__()

    def forward(self, x):
        return x

# The Identity class takes *args and **kwargs but doesn't use them
# We can pass any arguments and they will be accepted
ARGS = {
    'args': (),
    'kwargs': {}
}
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 Identity — 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 Identity 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
Wide_ResNet Ran
ran on a synthesized input · constructed an object; no result was computed
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("84de817b353f7c2c")
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, Wide_ResNet: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'depth': 10,
    'widen_factor': 2,
    'num_classes': 10,
    'input_channels': 3,
    'sum_pool': False,
    'norm': None,
    'leak': 0.2,
    'dropout_rate': 0.0
}
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 Wide_ResNet — 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 Wide_ResNet 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
conv3x3 Ran
ran; our contract draft was wrong, not the code
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("bf19b9ba510a9e4a")
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 Conv2d. 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 as t
import torch.nn as nn
import torch.nn.functional as tnnF
import torch.distributions as tdist

t.manual_seed(0)

ARGS = {
    'in_planes': 3,
    'out_planes': 64,
    'stride': 1,
}
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 Conv2d — a Conv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv2d 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
f_eq Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/contrib/scirl.py
code served (permissive licence) · get_code("7e084b093f24aee5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# The function f_eq takes two numeric arguments (a, b) and checks if their
# absolute difference is less than 1e-10. It works with floats or any numeric
# type that supports subtraction and comparison.

a = 0.5
b = 0.5 + 1e-11

ARGS = {
    "a": a,
    "b": b,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
f_geq Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/contrib/scirl.py
code served (permissive licence) · get_code("7613cdbabed93897")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# f_geq takes two numeric arguments and compares them
# It performs: a - b > -1e-10
# Both should be numeric types (int, float, or tensors)

a = 5.0
b = 4.9

ARGS = {
    "a": a,
    "b": b,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_linear_basis Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/core/basis.py
code served (permissive licence) · get_code("01397f68828e5085")
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

# Test case 1: include_action=False
ARGS = {
    "include_action": False
}

# Verify the function can be called with these arguments
if __name__ == "__main__":
    def get_linear_basis(include_action=False):
        def f(s, a):
            if include_action:
                sa = np.hstack((s, a))
                if len(sa.shape) == 1:
                    sa = np.expand_dims(sa, axis=0)
                return sa
            else:
                if len(s.shape) == 1:
                    s = np.expand_dims(s, axis=0)
                return s
        return f
    
    # Test with include_action=False
    basis_fn = get_linear_basis(**ARGS)
    s = np.array([1.0, 2.0, 3.0, 4.0])
    a = np.array([0.5, 0.6])
    result = basis_fn(s, a)
    assert result.shape == (1, 4), f"Expected shape (1, 4), got {result.shape}"
    
    # Test with include_action=True
    ARGS["include_action"] = True
    basis_fn_with_action = get_linear_basis(**ARGS)
    result_with_action = basis_fn_with_action(s, a)
    assert result_with_action.shape == (1, 6), f"Expected shape (1, 6), got {result_with_action.shape}"
    
    print("All tests passed!")

ARGS = {
    "include_action": 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 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
get_next_observation Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/main_sepsis.py
code served (permissive licence) · get_code("2fdd9847cb765877")
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

# Construct arguments for get_next_observation
# The function expects:
# - ob_list: numpy array of shape (N, ob_dim) containing observations
# - timestep_list: list/array of timestep values where 1 indicates episode boundary

N = 2  # batch size
ob_dim = 4  # observation dimension

# Create observation list as a numpy array
ob_list = np.array([
    [0.1, 0.2, 0.3, 0.4],
    [0.5, 0.6, 0.7, 0.8]
], dtype=np.float32)

# Create timestep list - values where 1 indicates episode boundary
# The function iterates through timestep_list[:-1] and checks if next_step == 1
# If next_step != 1, it copies ob_list[i+1] to ob_next_list[i]
timestep_list = np.array([0, 2, 1], dtype=np.int32)

ARGS = {
    "ob_list": ob_list,
    "timestep_list": timestep_list
}
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
get_norm Ran
ran; our contract draft was wrong, not the code
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("54752d9ecb18a6fb")
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 BatchNorm2d. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import types as _types

torch.manual_seed(0)

# Define the classes that get_norm might reference
class ActNorm(nn.Module):
    def __init__(self, in_channel, logdet=True):
        super().__init__()
        self.loc = nn.Parameter(torch.zeros(1, in_channel, 1, 1))
        self.scale = nn.Parameter(torch.ones(1, in_channel, 1, 1))
        self.register_buffer('initialized', torch.tensor(0, dtype=torch.uint8))
        self.logdet = logdet

    def initialize(self, input):
        with torch.no_grad():
            flatten = input.permute(1, 0, 2, 3).contiguous().view(input.shape[1], -1)
            mean = flatten.mean(1).unsqueeze(1).unsqueeze(2).unsqueeze(3).permute(1, 0, 2, 3)
            std = flatten.std(1).unsqueeze(1).unsqueeze(2).unsqueeze(3).permute(1, 0, 2, 3)
            self.loc.data.copy_(-mean)
            self.scale.data.copy_(1 / (std + 1e-06))

    def forward(self, input):
        _, _, height, width = input.shape
        if self.initialized.item() == 0:
            self.initialize(input)
            self.initialized.fill_(1)
        log_abs = lambda x: torch.log(torch.abs(x))
        log_abs_val = log_abs(self.scale)
        logdet = height * width * torch.sum(log_abs_val)
        if self.logdet:
            return (self.scale * (input + self.loc), logdet)
        else:
            return self.scale * (input + self.loc)

    def reverse(self, output):
        return output / self.scale - self.loc

class Identity(nn.Module):
    def __init__(self, *args, **kwargs):
        super().__init__()

    def forward(self, x):
        return x

class ConditionalActNorm(nn.Module):
    def __init__(self, *args, **kwargs):
        super().__init__()

class ConditionalInstanceNorm2dPlus(nn.Module):
    def __init__(self, *args, **kwargs):
        super().__init__()

class ContinuousConditionalActNorm(nn.Module):
    def __init__(self, *args, **kwargs):
        super().__init__()

logabs = lambda x: torch.log(torch.abs(x))

# Create the norms namespace
norms = _types.SimpleNamespace(**{
    n: globals()[n] for n in [
        'ActNorm', 'ConditionalActNorm', 'ConditionalInstanceNorm2dPlus', 
        'ContinuousConditionalActNorm', 'Identity', 'logabs'
    ] if n in globals()
})

def get_norm(n_filters, norm):
    if norm is None:
        return Identity()
    elif norm == 'batch':
        return nn.BatchNorm2d(n_filters, momentum=0.9)
    elif norm == 'instance':
        return nn.InstanceNorm2d(n_filters, affine=True)
    elif norm == 'layer':
        return nn.GroupNorm(1, n_filters)
    elif norm == 'act':
        return norms.ActNorm(n_filters, False)

# Test with different norm types
ARGS = {
    'n_filters': 4,
    'norm': 'batch'
}
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 BatchNorm2d — a BatchNorm2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a BatchNorm2d 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
new_reset Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/core/gym_env.py
code served (permissive licence) · get_code("72780861907b8bdd")
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 (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
from unittest.mock import MagicMock

class MockEnvironment:
    """Mock environment class that has a reset method"""
    def reset(self):
        return np.array([0.1, 0.2, 0.3, 0.4])

class Environment:
    """Concrete environment class with state attribute"""
    def __init__(self):
        self.state = None
    
    def reset(self):
        return np.array([0.5, 0.6, 0.7, 0.8])
    
    def new_reset(self, s_init=None):
        """The function being tested"""
        if s_init is None:
            self.state = self.__class__().reset()
        else:
            self.state = np.array(s_init)
        return self.state

# Create an instance to call the method on
env_instance = Environment()

# Define arguments for new_reset
ARGS = {
    "self": env_instance,
    "s_init": 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 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
new_timelimit_reset Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/core/gym_env.py
code served (permissive licence) · get_code("634baa09ecda13c4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import time

class StandInEnv:
    """Mock environment that implements reset() method."""
    def reset(self, s_init=None):
        # Return a simple observation
        return {"observation": torch.zeros(4)}

class StandInWrapper:
    """Mock wrapper class with the new_timelimit_reset method."""
    def __init__(self):
        self.env = StandInEnv()
        self._episode_started_at = None
        self._elapsed_steps = 0
    
    def new_timelimit_reset(self, s_init=None):
        import time
        self._episode_started_at = time.time()
        self._elapsed_steps = 0
        return self.env.reset(s_init)

# Create instance
wrapper = StandInWrapper()

ARGS = {
    "self": wrapper,
    "s_init": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
non_scalar_vectorize Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/src/contrib/scirl.py
code served (permissive licence) · get_code("59357864346147ca")
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
from numpy import prod, array

# Define a simple function that takes a 2D input and returns a 3D output
def simple_func(arg):
    """
    Takes input of shape (2,) and returns output of shape (10, 1)
    """
    # arg should have shape (2,)
    result = np.ones((10, 1)) * arg[0]
    return result

# Define input and output shapes
input_shape = (2,)
output_shape = (10, 1)

# Construct the ARGS dictionary
ARGS = {
    'func': simple_func,
    'input_shape': input_shape,
    'output_shape': output_shape
}
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 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
shape_el Ran
ran; our contract draft was wrong, not the code
dtak/batch-apprenticeship-learning/src/core/distributions.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from sihongho/robust_marl_with_state_uncertainty, is pointer only (licence: NONE) · get_code("aad5113d0b095522")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from sihongho/robust_marl_with_state_uncertainty, 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.

stack_observation Ran
ran on a synthesized input
dtak/batch-apprenticeship-learning/main_sepsis.py
code served (permissive licence) · get_code("c415e5b34587ae0f")
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 (5, 12) 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 valid arguments for stack_observation function

# Create ob_list: array of dimension (N, ob_dim)
# Using N=5 observations, ob_dim=3 observation dimension
ob_list = 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],
    [13.0, 14.0, 15.0]
], dtype=np.float32)

# Create timestep_list: indicates which timestep each observation is at
# Values range from 1 to stack_horizon
timestep_list = [1, 2, 3, 4, 4]

# stack_horizon: how many timesteps to stack together
stack_horizon = 4

ARGS = {
    "ob_list": ob_list,
    "timestep_list": timestep_list,
    "stack_horizon": stack_horizon
}
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 (5, 12) — shape (5, 12) 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
wide_basic Ran
ran on a synthesized input · constructed an object; no result was computed
wgrathwohl/JEM/train_wrn_ebm.py
code served (permissive licence) · get_code("8ecb413d596ba546")
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, wide_basic: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'in_planes': 4,
    'planes': 8,
    'dropout_rate': 0.1,
    'stride': 1,
    'norm': None,
    'leak': 0.2
}
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 wide_basic — 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 wide_basic 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
make_pdtype Not yet run
unverified (harvested, no recorded run)
dtak/batch-apprenticeship-learning/src/core/distributions.py
code served (permissive licence) · get_code("e86b3e45a52de298")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 4 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 3 later) and 13 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
conv3x3
brick of this paper's assembly: helper
74 other papers
17 dated earlier, 1 the same month, 56 later · 5 only as an assembly brick
1605.07146 (2016-05)
Wide Residual Networks
1707.01629 (2017-07)
Dual Path Networks
1802.05668 (2018-02)
Model compression via distillation and quantization
and 71 more not named here
shape_el 4 other papers
2 dated earlier, 2 later
1901.09216 (2019-01)
Modelling Bounded Rationality in Multi-Agent Interactions by Generalized Recursive Reasoning
1905.02363 (2019-05)
Dimension-Wise Importance Sampling Weight Clipping for Sample-Efficient Reinforcement Learning
2212.02705 (2022-12)
What is the Solution for State-Adversarial Multi-Agent Reinforcement Learning?
and 1 more not named here
ActNorm
brick of this paper's assembly: helper
1 other paper
1 dated earlier
1807.03039 (2018-07)
Glow: Generative Flow with Invertible 1x1 Convolutions
make_pdtype 1 other paper
1 later
2409.17411 (2024-09)
Enhancing Interpretability in Deep Reinforcement Learning through Semantic Clustering

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

Consider learning a policy purely on the basis of demonstrated behavior-that is, with no access to reinforcement signals, no knowledge of transition dynamics, and no further interaction with the environment. This strictly batch imitation learning problem arises wherever live experimentation is costly, such as in healthcare. One solution is simply to retrofit existing algorithms for apprenticeship learning to work in the offline setting. But such an approach leans heavily on off-policy evaluation or offline model estimation, and can be indirect and inefficient. We argue that a good solution should be able to explicitly parameterize a policy (i.e. respecting action conditionals), implicitly learn from rollout dynamics (i.e. leveraging state marginals), and-crucially-operate in an entirely offline fashion. To address this challenge, we propose a novel technique by energy-based distribution matching (EDM): By identifying parameterizations of the (discriminative) model of a policy with the (generative) energy function for state distributions, EDM yields a simple but effective solution that equivalently minimizes a divergence between the occupancy measure for the demonstrator and a model thereof for the imitator. Through experiments with application to control and healthcare settings, we illustrate consistent performance gains over existing algorithms for strictly batch imitation learning.

For agents

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

get_harvested_code_for_paper("2006.14154")
get_code_for_paper("2006.14154")
have("2006.14154")

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

curl https://syntology.ai/api/ran/2006.14154.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.14154.svg)](https://syntology.ai/paper/2006.14154)

Connect an agent — have() is free.