SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.07414 · NeurIPS · 2020

Combinatorial 3D Shape Generation via Sequential Assembly

Jungtaek Kim, Hyunsoo Chung, Minsu Cho, Jaesik Park

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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 4 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/2004.07414.json · badge: /api/ran/2004.07414.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
POSTECH-CVLab/Combinatorial-3D-Shape-Generation canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
FunctionStatusWhere it lives
random_all Ran
ran on a synthesized input
POSTECH-CVLab/Combinatorial-3D-Shape-Generation/src_generation/wrappers.py
code served (permissive licence) · get_code("b88b3b7fc625c015")
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 MockBricks. 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 copy
import time

class MockBricks:
    """Stand-in class for bricks_initial that supports sample() and add() methods."""
    
    def __init__(self):
        self.items = [
            torch.randn(2, 4),
            torch.randn(2, 4),
            torch.randn(2, 4),
            torch.randn(2, 4),
            torch.randn(2, 4),
        ]
        self.index = 0
    
    def sample(self):
        """Return a list with one sampled item."""
        item = self.items[self.index % len(self.items)]
        self.index += 1
        return [item]
    
    def add(self, item):
        """Add an item to the collection."""
        self.items.append(item)

torch.manual_seed(0)

bricks_initial = MockBricks()
num_bricks = 5

ARGS = {
    'bricks_initial': bricks_initial,
    'num_bricks': num_bricks,
}
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 MockBricks — a MockBricks output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockBricks 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
random_eval_all Ran
ran on a synthesized input
POSTECH-CVLab/Combinatorial-3D-Shape-Generation/src_generation/wrappers.py
code served (permissive licence) · get_code("b5a9dd3d7ec08fc3")
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 BricksContainer. 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
import copy

class BricksContainer:
    """Stand-in class for a container that holds bricks and supports sampling and adding."""
    
    def __init__(self, initial_bricks=None):
        """Initialize with a list of bricks."""
        if initial_bricks is None:
            initial_bricks = []
        self.bricks = list(initial_bricks)
    
    def sample(self, num_samples):
        """Sample num_samples bricks from available bricks."""
        # Return a list of sampled brick indices or identifiers
        num_available = len(self.bricks)
        if num_available == 0:
            # Generate candidate bricks if none exist yet
            return [{'id': i, 'value': np.random.rand()} for i in range(num_samples)]
        # Sample from existing bricks with replacement
        indices = np.random.choice(num_available, min(num_samples, num_available), replace=True)
        return [self.bricks[i] for i in indices]
    
    def add(self, brick):
        """Add a brick to the container."""
        self.bricks.append(brick)
    
    def __deepcopy__(self, memo):
        """Support deep copying."""
        new_bricks = copy.deepcopy(self.bricks, memo)
        new_container = BricksContainer(new_bricks)
        return new_container


def evaluation_function(bricks_container):
    """Simple evaluation function that returns a score based on number of bricks."""
    # Lower score is better (based on argsort()[0] logic in the function)
    return float(len(bricks_container.bricks)) + np.random.rand() * 0.1


# Set up initial bricks container with a few starter bricks
initial_bricks = [
    {'id': 0, 'value': 0.5},
    {'id': 1, 'value': 0.7}
]
bricks_initial = BricksContainer(initial_bricks)

ARGS = {
    'bricks_initial': bricks_initial,
    'num_bricks': 4,
    'num_acq': 3,
    'fun_evaluation': evaluation_function
}
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 BricksContainer — a BricksContainer output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a BricksContainer 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
score_bricks_axis Ran
ran on a synthesized input
POSTECH-CVLab/Combinatorial-3D-Shape-Generation/src_generation/maximize_explicit_scores.py
code served (permissive licence) · get_code("fa81b0c2999800af")
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 float32, 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 numpy as np

class Brick:
    def __init__(self, vertices):
        self.vertices = vertices
    
    def get_vertices(self):
        return self.vertices

class Bricks:
    def __init__(self, bricks_list):
        self.bricks_list = bricks_list
    
    def get_bricks(self):
        return self.bricks_list

# Create sample vertices for bricks
# Using distinct dimensions: 5 vertices (rows) × 3 coordinates (cols for x, y, z)
vertices_1 = np.array([
    [1.0, 2.0, 3.0],
    [2.0, 3.0, 4.0],
    [3.0, 4.0, 5.0],
    [4.0, 5.0, 6.0],
    [5.0, 6.0, 7.0]
], dtype=np.float32)

vertices_2 = np.array([
    [0.5, 1.5, 2.5],
    [1.5, 2.5, 3.5],
    [2.5, 3.5, 4.5],
    [3.5, 4.5, 5.5],
    [4.5, 5.5, 6.5]
], dtype=np.float32)

brick_1 = Brick(vertices_1)
brick_2 = Brick(vertices_2)

bricks_container = Bricks([brick_1, brick_2])

ARGS = {
    "bricks_": bricks_container,
    "ind_axis": 0
}
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 float32, 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
score_bricks_height Ran
ran on a synthesized input
POSTECH-CVLab/Combinatorial-3D-Shape-Generation/src_generation/maximize_explicit_scores.py
code served (permissive licence) · get_code("40f71b7bbda0dbc2")
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 torch
import numpy as np

class Brick:
    def __init__(self, position, height):
        self.position = position
        self.height = height
    
    def get_position(self):
        return self.position

class Bricks:
    def __init__(self, bricks_list):
        self.bricks_list = bricks_list
    
    def get_bricks(self):
        return self.bricks_list

torch.manual_seed(0)

brick1 = Brick(position=[0.0, 0.0, 1.0], height=2.5)
brick2 = Brick(position=[1.0, 1.0, 2.0], height=1.5)
brick3 = Brick(position=[2.0, 2.0, 0.5], height=3.0)

bricks_container = Bricks([brick1, brick2, brick3])

ARGS = {
    "bricks_": bricks_container
}
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 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

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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 2026-11-21.

Repositories linked to this paper

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

Abstract

Sequential assembly with geometric primitives has drawn attention in robotics and 3D vision since it yields a practical blueprint to construct a target shape. However, due to its combinatorial property, a greedy method falls short of generating a sequence of volumetric primitives. To alleviate this consequence induced by a huge number of feasible combinations, we propose a combinatorial 3D shape generation framework. The proposed framework reflects an important aspect of human generation processes in real life -we often create a 3D shape by sequentially assembling unit primitives with geometric constraints. To find the desired combination regarding combination evaluations, we adopt Bayesian optimization, which is able to exploit and explore efficiently the feasible regions constrained by the current primitive placements. An evaluation function conveys global structure guidance for an assembly process and stability in terms of gravity and external forces simultaneously. Experimental results demonstrate that our method successfully generates combinatorial 3D shapes and simulates more realistic generation processes. We also introduce a new dataset for combinatorial 3D shape generation. All the codes are available at https: //github.com/POSTECH-CVLab/Combinatorial-3D-Shape-Generation.

For agents

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

get_harvested_code_for_paper("2004.07414")
get_code_for_paper("2004.07414")
have("2004.07414")

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

curl https://syntology.ai/api/ran/2004.07414.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.07414.svg)](https://syntology.ai/paper/2004.07414)

Connect an agent — have() is free.