SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2301.06387 · 2023

PECAN: Leveraging Policy Ensemble for Context-Aware Zero-Shot Human-AI Coordination

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 2 failures of our instrument; 2 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/2301.06387.json · badge: /api/ran/2301.06387.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
LxzGordon/pecan_human_AI_coordination canonical (classifier label)
classifier basis not recorded on the link
5 ran · 2 no recorded run
FunctionStatusWhere it lives
cross_entropy Ran
ran on a synthesized input
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/human_aware_rl/utils.py
code served (permissive licence) · get_code("9ed8cfc43ed32bfd")
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

# Create action_probs: probability distribution over classes for each example
# Shape: (num_examples, num_classes) = (2, 4)
action_probs = np.array([
    [0.1, 0.6, 0.2, 0.1],
    [0.3, 0.2, 0.4, 0.1]
], dtype=float)

# Create y: class labels for each example (not one-hot encoded)
# Shape: (num_examples,) = (2,)
# Values must be valid indices into the class dimension (0-3)
y = np.array([1, 2], dtype=int)

# eps parameter with default value
eps = 0.0001

ARGS = {
    "action_probs": action_probs,
    "y": y,
    "eps": eps
}
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
get_max_iter Ran
ran on a synthesized input
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/human_aware_rl/utils.py
code served (permissive licence) · get_code("3a08670f59f30530")
How it ran

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

import os
import tempfile
import shutil

# Create a temporary directory with mock PBT iteration folders
temp_dir = tempfile.mkdtemp()

# Create some mock PBT iteration folders with trailing numbers
os.makedirs(os.path.join(temp_dir, 'pbt_iter_0'))
os.makedirs(os.path.join(temp_dir, 'pbt_iter_5'))
os.makedirs(os.path.join(temp_dir, 'pbt_iter_12'))
os.makedirs(os.path.join(temp_dir, 'pbt_iter_3'))

ARGS = {
    'agent_folder': temp_dir
}
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 int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_trailing_number Ran
ran on a synthesized input
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/human_aware_rl/utils.py
code served (permissive licence) · get_code("206b32402a1b9e7a")
How it ran

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

import re

def get_trailing_number(s):
    """
    Get the trailing number from a string,
    i.e. 'file123' -> '123'
    """
    m = re.search('\\d+$', s)
    return int(m.group()) if m else None

ARGS = {
    "s": "file123"
}
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 int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_output_format Ran
ran; our contract draft was wrong, not the code
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/baselines/baselines/logger.py
pointer only (licence: MIT); the stored copy get_code serves, from TonghanWang/EITI-EDTI, is pointer only (licence: MIT) · get_code("817d5caf0e3d3706")
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 TonghanWang/EITI-EDTI, and MIT, as GitHub detects it on that repository, is in the permissive set, but a licence decision keeps the body a pointer: its licence file names an outside party, so it may cover vendored code rather than the repository owner's files (recorded as pointer:licence_names_third_party_corporation). 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.

register Ran
ran; our contract draft was wrong, not the code
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/baselines/baselines/common/models.py
code served (permissive licence) · get_code("a44574392d168384")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) 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 torch
import numpy as np

# The register function is a decorator factory that takes a name parameter
# and returns a decorator that registers functions in a mapping dict.
# We just need to provide a string name.

ARGS = {
    "name": "my_function"
}
Setup block drafted by
a language model, Claude Sonnet 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 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
profile Not yet run
unverified (harvested, no recorded run)
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/baselines/baselines/logger.py
pointer only (licence: MIT); the stored copy get_code serves, from TonghanWang/EITI-EDTI, is pointer only (licence: MIT) · get_code("8455885a0d199802")
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 TonghanWang/EITI-EDTI, and MIT, as GitHub detects it on that repository, is in the permissive set, but a licence decision keeps the body a pointer: its licence file names an outside party, so it may cover vendored code rather than the repository owner's files (recorded as pointer:licence_names_third_party_corporation). 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.

read_json Not yet run
unverified (harvested, no recorded run)
LxzGordon/pecan_human_AI_coordination/overcookedgym/human_aware_rl/baselines/baselines/logger.py
code served (permissive licence) · get_code("88ef5a17fc986d9a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 6 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 2 the same month, 6 later) and 1 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
make_output_format 73 other papers
35 dated earlier, 2 the same month, 36 later
1611.05763 (2016-11)
Learning to reinforcement learn
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
1806.02473 (2018-06)
Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
and 70 more not named here
profile 61 other papers
22 dated earlier, 2 the same month, 36 later, 1 same year, order not determinable
1611.05763 (2016-11)
Learning to reinforcement learn
1905.06750 (2019-05)
Random Expert Distillation: Imitation Learning via Expert Policy Support Estimation
1906.05841 (2019-06)
Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards
and 58 more not named here
read_json 22 other papers
18 dated earlier, 4 later
1611.05763 (2016-11)
Learning to reinforcement learn
1711.00123 (2017-11)
Backpropagation through the Void: Optimizing control variates for black-box gradient estimation
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
and 19 more not named here
register 12 other papers
11 dated earlier, 1 later
1611.05763 (2016-11)
Learning to reinforcement learn
1708.05144 (2017-08)
Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation
1905.02363 (2019-05)
Dimension-Wise Importance Sampling Weight Clipping for Sample-Efficient Reinforcement Learning
and 9 more not named here
get_max_iter 1 other paper
1 later
2306.03034 (2023-06)
Tackling Cooperative Incompatibility for Zero-Shot Human-AI Coordination
get_trailing_number 1 other paper
1 later
2306.03034 (2023-06)
Tackling Cooperative Incompatibility for Zero-Shot Human-AI Coordination

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

Zero-shot human-AI coordination holds the promise of collaborating with humans without human data. Prevailing methods try to train the ego agent with a population of partners via self-play. However, these methods suffer from two problems: 1) The diversity of a population with finite partners is limited, thereby limiting the capacity of the trained ego agent to collaborate with a novel human; 2) Current methods only provide a common best response for every partner in the population, which may result in poor zero-shot coordination performance with a novel partner or humans. To address these issues, we first propose the policy ensemble method to increase the diversity of partners in the population, and then develop a context-aware method enabling the ego agent to analyze and identify the partner's potential policy primitives so that it can take different actions accordingly. In this way, the ego agent is able to learn more universal cooperative behaviors for collaborating with diverse partners. We conduct experiments on the Overcooked environment, and evaluate the zero-shot human-AI coordination performance of our method with both behavior-cloned human proxies and real humans. The results demonstrate that our method significantly increases the diversity of partners and enables ego agents to learn more diverse behaviors than baselines, thus achieving state-of-the-art performance in all scenarios. We also open-source a human-AI coordination study framework on the Overcooked for the convenience of future studies.

For agents

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

get_harvested_code_for_paper("2301.06387")
get_code_for_paper("2301.06387")
have("2301.06387")

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

curl https://syntology.ai/api/ran/2301.06387.json

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

[![Syntology run record](https://syntology.ai/api/ran/2301.06387.svg)](https://syntology.ai/paper/2301.06387)

Connect an agent — have() is free.