SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1812.09755 · 2018

Learning when to Communicate at Scale in Multiagent Cooperative and Competitive Tasks

arXiv · PDF · Open in the Atlas

Code that ran

9 of 11 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, 0 violated it and 8 have no contract verdict; 0 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/1812.09755.json · badge: /api/ran/1812.09755.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
IC3Net/IC3Net canonical (classifier label)
classifier basis: census
5 ran · 2 no recorded run
apsdehal/gym-starcraft community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
RL-WFU/IC3Net-Highway reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
get_add_mat Ran
ran on a synthesized input
IC3Net/IC3Net/ic3net-envs/ic3net_envs/traffic_helper.py
code served (permissive licence) · get_code("66f904cd49b5f2bb")
How it ran

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

import numpy as np

# Create a small grid for testing
# The function expects a 2D numpy array that it will copy and modify
grid = np.zeros((9, 12), dtype=int)

ARGS = {
    'dims': (9, 12),
    'grid': grid,
    'difficulty': 'medium'
}
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_closest Ran
ran on a synthesized input
apsdehal/gym-starcraft/gym_starcraft/utils.py
code served (permissive licence) · get_code("b0f41d218fa1d86d")
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 Unit. 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 math

class Unit:
    def __init__(self, x, y):
        self.x = x
        self.y = y

class State:
    def __init__(self, units):
        self.units = units

# Create some units for player 0
unit1 = Unit(0, 0)
unit2 = Unit(3, 4)
unit3 = Unit(1, 1)

# Create some units for player 1
opp_unit1 = Unit(5, 5)
opp_unit2 = Unit(2, 2)
opp_unit3 = Unit(10, 10)

# Create state with units indexed by player_id
state = State({
    0: [unit1, unit2, unit3],
    1: [opp_unit1, opp_unit2, opp_unit3]
})

# The unit we want to find the closest opponent to
unit = Unit(1, 0)

player_id = 1

ARGS = {
    "unit": unit,
    "state": state,
    "player_id": player_id
}
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 Unit — a Unit output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Unit 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_distance Ran
ran on a synthesized input
apsdehal/gym-starcraft/gym_starcraft/utils.py
code served (permissive licence) · get_code("4cae8240fd6f06df")
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 math

def get_distance(x1, y1, x2, y2):
    return math.hypot(x2 - x1, y2 - y1)

ARGS = {
    "x1": 0.0,
    "y1": 0.0,
    "x2": 3.0,
    "y2": 4.0,
}
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 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
get_road_blocks Ran
ran on a synthesized input
IC3Net/IC3Net/ic3net-envs/ic3net_envs/traffic_helper.py
code served (permissive licence) · get_code("4a6659e981dad238")
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 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

# The function get_road_blocks takes three parameters:
# w: width (integer)
# h: height (integer)
# difficulty: string key into the road_blocks dict ('easy', 'medium', or 'hard')

ARGS = {
    'w': 20,
    'h': 15,
    'difficulty': 'easy'
}
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 list, ndim 2, 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_weakest Ran
ran on a synthesized input
apsdehal/gym-starcraft/gym_starcraft/utils.py
code served (permissive licence) · get_code("203cd54da11b7d04")
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 Unit. 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 math

class Unit:
    def __init__(self, health, shield):
        self.health = health
        self.shield = shield

class State:
    def __init__(self, units):
        self.units = units

# Create some units with different health + shield values
unit1 = Unit(health=10, shield=5)  # total: 15
unit2 = Unit(health=3, shield=2)   # total: 5 (weakest)
unit3 = Unit(health=8, shield=8)   # total: 16

# Create opponent units list
opponent_units = [unit1, unit2, unit3]

# Create state with units indexed by player_id
state = State(units={1: opponent_units})

# Create a unit to pass as first argument (not used in function body, but required)
unit = Unit(health=20, shield=10)

ARGS = {
    "unit": unit,
    "state": state,
    "player_id": 1
}
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 Unit — a Unit output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Unit 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
goal_reached Ran
ran on a synthesized input
IC3Net/IC3Net/ic3net-envs/ic3net_envs/traffic_helper.py
code served (permissive licence) · get_code("2c15fd3b782cb39c")
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)

# Analyzing the function:
# - place_i: appears to be an integer index
# - curr: appears to be a value that will be checked for membership in finish_points
# - finish_points: appears to be a sequence (list, tuple, or tensor) that can be sliced

# The function checks if curr is in the concatenation of two slices of finish_points
# So finish_points should be a list or sequence of comparable values

# Let's create simple integer values for this
place_i = 1
curr = 5
finish_points = [3, 5, 7, 9, 11]

ARGS = {
    "place_i": place_i,
    "curr": curr,
    "finish_points": finish_points,
}
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
nPr Ran
ran, honoured the contract we drafted
RL-WFU/IC3Net-Highway/ic3net-envs/ic3net_envs/traffic_junction_env.py
code served (permissive licence) · get_code("37a75d30add374c0")
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 math

def nPr(n, r):
    f = math.factorial
    return f(n) // f(n - r)

ARGS = {
    "n": 5,
    "r": 3
}
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
read_file Ran
ran on a synthesized input
IC3Net/IC3Net/plot_script.py
code served (permissive licence) · get_code("a9fef680e2912a66")
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 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 tempfile
import os

# Create a temporary file with test data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("Epoch 1\tSome header\n")
temp_file.write("metric_value 42.5\ttest_term [ 1.0 2.0 3.0 ]\n")
temp_file.write("Epoch 2\tSome header\n")
temp_file.write("metric_value 45.3\ttest_term [ 2.0 3.0 4.0 ]\n")
temp_file.close()

ARGS = {
    'vec': [],
    'file_name': temp_file.name,
    'scalar': False,
    'term': 'test_term'
}
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 list, ndim 2, 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
select_action Ran
ran on a synthesized input
IC3Net/IC3Net/action_utils.py
code served (permissive licence) · get_code("dbf148c04f6bc762")
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 (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 torch
import torch.nn as nn

torch.manual_seed(0)

# Test case 1: continuous action selection
class ContinuousArgs:
    def __init__(self):
        self.continuous = True

args_continuous = ContinuousArgs()
action_mean = torch.randn(2, 4)
action_std = torch.ones(2, 4)
action_log_std = torch.zeros(2, 4)
action_out_continuous = (action_mean, action_log_std, action_std)

ARGS = {
    "args": args_continuous,
    "action_out": action_out_continuous
}
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 (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
init Not yet run
unverified (harvested, no recorded run)
IC3Net/IC3Net/data.py
code served (permissive licence) · get_code("8cab58f1257e1400")
How it ran

no invocation recorded.

translate_action Not yet run
unverified (harvested, no recorded run)
IC3Net/IC3Net/action_utils.py
code served (permissive licence) · get_code("420dd86fe5989706")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 11 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

Learning when to communicate and doing that effectively is essential in multi-agent tasks. Recent works show that continuous communication allows efficient training with back-propagation in multi-agent scenarios, but have been restricted to fully-cooperative tasks. In this paper, we present Individualized Controlled Continuous Communication Model (IC3Net) which has better training efficiency than simple continuous communication model, and can be applied to semi-cooperative and competitive settings along with the cooperative settings. IC3Net controls continuous communication with a gating mechanism and uses individualized rewards foreach agent to gain better performance and scalability while fixing credit assignment issues. Using variety of tasks including StarCraft BroodWars explore and combat scenarios, we show that our network yields improved performance and convergence rates than the baselines as the scale increases. Our results convey that IC3Net agents learn when to communicate based on the scenario and profitability.

For agents

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

get_harvested_code_for_paper("1812.09755")
get_code_for_paper("1812.09755")
have("1812.09755")

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

curl https://syntology.ai/api/ran/1812.09755.json

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

[![Syntology run record](https://syntology.ai/api/ran/1812.09755.svg)](https://syntology.ai/paper/1812.09755)

Connect an agent — have() is free.