SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2302.05103 · ICML · 2023

Controllability-Aware Unsupervised Skill Discovery

Pieter Abbeel, Kimin Lee, Seohong Park, Youngwoon Lee

arXiv · PDF · Open in the Atlas

Code that ran

7 of 10 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; 2 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 4 failures of our instrument; 3 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/2302.05103.json · badge: /api/ran/2302.05103.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
seohongpark/csd-manipulation canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
seohongpark/csd-locomotion canonical (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
artberryx/csd-public marked official in Papers with Code
Papers with Code archive
3 ran · 3 no recorded run
FunctionStatusWhere it lives
colorize Ran
ran; our contract draft was wrong, not the code
seohongpark/csd-locomotion/dowel/utils.py
code served (permissive licence) · get_code("68a57588acd1e63a")
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 str. 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

color2num = dict(gray=30, red=31, green=32, yellow=33, blue=34, magenta=35, cyan=36, white=37, crimson=38)

def colorize(string, color, bold=False, highlight=False):
    """Colorize the string for console output."""
    attr = []
    num = color2num[color]
    if highlight:
        num += 10
    attr.append(str(num))
    if bold:
        attr.append('1')
    return '\x1b[%sm%s\x1b[0m' % (';'.join(attr), string)


ARGS = {
    'string': 'Hello World',
    'color': 'blue',
    'bold': True,
    'highlight': False
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_2d_colors Ran
ran, honoured the contract we drafted
seohongpark/csd-manipulation/utils.py
code served (permissive licence) · get_code("ea62c621866571a4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned ndarray, ndim 2, dtype family float, shape (3, 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 valid arguments for get_2d_colors
# The function expects:
# - points: array-like of shape (N, 2) with 2D coordinates
# - min_point: array-like of shape (2,) with minimum bounds
# - max_point: array-like of shape (2,) with maximum bounds

points = np.array([
    [0.5, 0.3],
    [1.2, 0.8],
    [0.1, 0.9]
])

min_point = np.array([0.0, 0.0])

max_point = np.array([2.0, 1.0])

ARGS = {
    "points": points,
    "min_point": min_point,
    "max_point": max_point
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 (3, 4) — shape (3, 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_option_colors Ran
ran, honoured the contract we drafted
seohongpark/csd-manipulation/utils.py
code served (permissive licence) · get_code("eacc7920f6eceaf0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned ndarray, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small options array with shape (num_options, dim_option)
# Using distinct sizes: 2 options, 4 dimensions
options = np.random.randn(2, 4)

ARGS = {
    "options": options
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, sklearn — 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
l2_norm Ran
ran, violated the contract we drafted
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("e47eba06f6824dbd")
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 EagerTensor, 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 tensorflow as tf
import numpy as np

def l2_norm(v, eps=1e-12):
    return v / (tf.reduce_sum(input_tensor=v ** 2) ** 0.5 + eps)

# Create a small TensorFlow tensor for testing
v = tf.constant([[1.0, 2.0, 3.0, 4.0],
                 [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32)

eps = 1e-12

ARGS = {
    "v": v,
    "eps": eps
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, 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
setup_evaluation Ran
ran; our contract draft was wrong, not the code
seohongpark/csd-manipulation/utils.py
code served (permissive licence) · get_code("16ebb9997d99bab8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned 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

ARGS = {
    "num_eval_options": 5,
    "num_eval_trajectories_per_option": 2,
    "num_skills": 2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, sklearn — 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
spectral_norm Ran
ran; our contract draft was wrong, not the code
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("ce82cd8ce37faf68")
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 SymbolicTensor. 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 tensorflow as tf
import numpy as np

# Enable TensorFlow v1 compatibility
tf.compat.v1.disable_eager_execution()

# Create a simple weight tensor
# w_shape will be [height, width, ..., output_dim]
# For testing, use shape [2, 3, 4] so w_shape[-1] = 4
weight_value = np.random.randn(2, 3, 4).astype(np.float32)
w = tf.constant(weight_value)

# Create a TensorFlow session and initialize variables
sess = tf.compat.v1.Session()

# Create a placeholder for is_training
is_training = tf.compat.v1.placeholder(tf.bool, shape=())

ARGS = {
    'w': w,
    'iteration': 1,
    'name': 'test',
    'is_training': is_training
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
store_args Ran
ran; our contract draft was wrong, not the code
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("d123b177811d57b3")
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 torch
import inspect
import functools

torch.manual_seed(0)

class SimpleClass:
    def __init__(self):
        self.value = 42
    
    def simple_method(self, x, y=10, z=20):
        """A simple method to decorate with store_args"""
        return x + y + z

def setup_store_args_test():
    """Set up test instance and arguments"""
    instance = SimpleClass()
    return instance

ARGS = {
    "method": SimpleClass.simple_method
}
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 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
Discriminator Not yet run
unverified (harvested, no recorded run)
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("6791c540f021f7a6")
How it ran

no invocation recorded.

fully_conneted Not yet run
unverified (harvested, no recorded run)
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("e51948cd326a355e")
How it ran

no invocation recorded.

snn Not yet run
unverified (harvested, no recorded run)
artberryx/csd-public/baselines/her/discriminator.py
code served (permissive licence) · get_code("fa9fb1b4a96d224b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 2 later) and 8 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

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

FunctionAlso present inEarliest-datedOthers named
colorize 17 other papers
7 dated earlier, 1 the same month, 8 later, 1 same year, order not determinable
1802.09477 (2018-02)
Addressing Function Approximation Error in Actor-Critic Methods
1909.12255 (2019-09)
Harnessing Structures for Value-Based Planning and Reinforcement Learning
2006.07442 (2020-06)
Self-Imitation Learning via Generalized Lower Bound Q-learning
and 14 more not named here
store_args
brick of this paper's assembly: helper
4 other papers
2 dated earlier, 2 later
2005.03420 (2020-05)
Curious Hierarchical Actor-Critic Reinforcement Learning
2301.10034 (2023-01)
Open-World Multi-Task Control Through Goal-Aware Representation Learning and Adaptive Horizon Prediction
2312.02405 (2023-12)
BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks
and 1 more not named here

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

One of the key capabilities of intelligent agents is the ability to discover useful skills without external supervision. However, the current unsupervised skill discovery methods are often limited to acquiring simple, easy-to-learn skills due to the lack of incentives to discover more complex, challenging behaviors. We introduce a novel unsupervised skill discovery method, Controllabilityaware Skill Discovery (CSD), which actively seeks complex, hard-to-control skills without supervision. The key component of CSD is a controllability-aware distance function, which assigns larger values to state transitions that are harder to achieve with the current skills. Combined with distance-maximizing skill discovery, CSD progressively learns more challenging skills over the course of training as our jointly trained distance function reduces rewards for easy-toachieve skills. Our experimental results in six robotic manipulation and locomotion environments demonstrate that CSD can discover diverse complex skills including object manipulation and locomotion skills with no supervision, significantly outperforming prior unsupervised skill discovery methods. Videos and code are available at https://seohong.me/projects/csd/

For agents

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

get_harvested_code_for_paper("2302.05103")
get_code_for_paper("2302.05103")
have("2302.05103")

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

curl https://syntology.ai/api/ran/2302.05103.json

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

[![Syntology run record](https://syntology.ai/api/ran/2302.05103.svg)](https://syntology.ai/paper/2302.05103)

Connect an agent — have() is free.