SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.17842 · CVPR · 2023

Shepherding Slots to Objects: Towards Stable and Robust Object-Centric Learning

Seon Kim, Ho-Jin Choi, Jinwoo Kim, Janghyuk Choi

arXiv · PDF · Open in the Atlas

Code that ran

8 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (7 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 1 failure 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/2303.17842.json · badge: /api/ran/2303.17842.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
object-understanding/slash marked official in Papers with Code
Papers with Code archive
8 ran · 1 no recorded run
FunctionStatusWhere it lives
Decoder Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("915a1c640fbf2324")
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, Decoder: 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 numpy as np
from torch import nn

torch.manual_seed(0)

class MockConfig:
    def __init__(self):
        self.MODEL = MockModel()
        self.DATA = MockData()

class MockModel:
    def __init__(self):
        self.SLOT = MockSlot()
        self.DEC_HID_DIM = 8
        self.HID_DIM = 16
        self.DEC_INIT_SIZE = 4
        self.DEC_DEPTH = 4

class MockSlot:
    def __init__(self):
        self.DIM = 4

class MockData:
    def __init__(self):
        self.IMG_SIZE = 16

cfg = MockConfig()

ARGS = {
    "cfg": cfg
}
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 Decoder — 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 Decoder 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
Encoder Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("914d2eb710d4bb7e")
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, Encoder: 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
import numpy as np

torch.manual_seed(0)

class Config:
    def __init__(self):
        self.MODEL = ModelConfig()
        self.DATA = DataConfig()

class ModelConfig:
    def __init__(self):
        self.ENC_DEPTH = 4
        self.HID_DIM = 8

class DataConfig:
    def __init__(self):
        self.IMG_SIZE = 4

cfg = Config()

ARGS = {
    'cfg': cfg
}
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 Encoder — 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 Encoder 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
PositionEncoder Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("e821b1827f3a6e73")
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, PositionEncoder: 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)

class MockConfig:
    class MODEL:
        class SLOT:
            DIM = 64
    class WEAK_SUP:
        TYPE = 'bbox'

cfg = MockConfig()

ARGS = {
    'cfg': cfg
}
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 PositionEncoder — 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 PositionEncoder 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
PositionPredictor Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("d0744a8714a58f6f")
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, PositionPredictor: 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)

class Config:
    def __init__(self):
        self.WEAK_SUP = WeakSupConfig()
        self.MODEL = ModelConfig()
        self.POS_PRED = PosPredConfig()

class WeakSupConfig:
    def __init__(self):
        self.TYPE = 'bbox'

class ModelConfig:
    def __init__(self):
        self.SLOT = SlotConfig()

class SlotConfig:
    def __init__(self):
        self.DIM = 64

class PosPredConfig:
    def __init__(self):
        self.PP_SIZE = 'base'

cfg = Config()

ARGS = {
    'cfg': cfg
}
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 PositionPredictor — 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 PositionPredictor 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
SlotAttention Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("77808f7557795579")
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, SlotAttention: 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
import numpy as np
from collections import namedtuple

torch.manual_seed(0)

# Create a minimal config object with all required attributes
class MinimalConfig:
    def __init__(self):
        self.MODEL = namedtuple('MODEL', ['SLOT', 'HID_DIM', 'MLP_HID_DIM'])(
            SLOT=namedtuple('SLOT', ['NUM', 'DIM', 'ITERATIONS', 'ATTN_HEADS', 'TEMPERATURE', 'ATTN_SMOOTH', 'ATTN_SMOOTH_SIZE'])(
                NUM=4,
                DIM=16,
                ITERATIONS=3,
                ATTN_HEADS=2,
                TEMPERATURE=1.0,
                ATTN_SMOOTH='',
                ATTN_SMOOTH_SIZE=3
            ),
            HID_DIM=8,
            MLP_HID_DIM=12
        )
        self.WEAK_SUP = namedtuple('WEAK_SUP', ['TYPE', 'INIT_USING_SUP', 'SPLIT'])(
            TYPE='',
            INIT_USING_SUP=False,
            SPLIT=namedtuple('SPLIT', ['RATIO', 'TRAIN'])(
                RATIO=0.5,
                TRAIN=namedtuple('TRAIN', ['BATCH_FUSION', 'BATCH_FUSION_RATIO'])(
                    BATCH_FUSION=False,
                    BATCH_FUSION_RATIO=0.5
                )
            )
        )
        self.POS_PRED = namedtuple('POS_PRED', ['USE_POS_PRED', 'USE_GT', 'PP_SIZE'])(
            USE_POS_PRED=False,
            USE_GT=False,
            PP_SIZE='base'
        )
        self.TRAIN = namedtuple('TRAIN', ['BATCH_SIZE'])(
            BATCH_SIZE=4
        )

cfg = MinimalConfig()

ARGS = {
    'cfg': cfg,
    'eps': 1e-8
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SlotAttention — 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 SlotAttention 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
SoftPositionEmbed Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("0d768f1463738a88")
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, SoftPositionEmbed: 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 numpy as np
from torch import nn

torch.manual_seed(0)

ARGS = {
    "hidden_size": 8,
    "resolution": (4, 3),
}
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 SoftPositionEmbed — 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 SoftPositionEmbed 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
WNConv Ran
ran on a synthesized input · constructed an object; no result was computed
object-understanding/slash/model.py
code served (permissive licence) · get_code("c26a57f28ea60d59")
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, WNConv: 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)

# Create a small input tensor for testing
# Shape: (batch=2, in_channels=3, height=4, width=4)
input_tensor = torch.randn(2, 3, 4, 4)

# Define ARGS mapping all parameters of WNConv.__init__
ARGS = {
    'in_channels': 3,
    'out_channels': 5,
    'kernel_size': 3,
    'padding': 1,
    'bias': False,
    'norm_type': 'softmax',
    'use_normed_logits': False,
}
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 WNConv — 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 WNConv 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
build_grid Ran
ran; our contract draft was wrong, not the code
object-understanding/slash/model.py
code served (permissive licence) · get_code("58fa0de4f2f3e646")
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 4, dtype family float, shape (1, 4, 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
import torch

torch.manual_seed(0)

ARGS = {
    "resolution": [4, 3]
}
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 Tensor, ndim 4, dtype family float, shape (1, 4, 3, 4) — shape (1, 4, 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
SlotAttentionAutoEncoder Not yet run
unverified (harvested, no recorded run)
object-understanding/slash/model.py
code served (permissive licence) · get_code("40399a09da048be6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 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
build_grid
brick of this paper's assembly: helper
2 other papers
1 dated earlier, 1 later
2209.11924 (2022-09)
Interventional Causal Representation Learning
2401.10148 (2024-01)
Explicitly Disentangled Representations in Object-Centric Learning

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

Object-centric learning (OCL) aspires general and compositional understanding of scenes by representing a scene as a collection of object-centric representations. OCL has also been extended to multi-view image and video datasets to apply various data-driven inductive biases by utilizing geometric or temporal information in the multi-image data. Single-view images carry less information about how to disentangle a given scene than videos or multi-view images do. Hence, owing to the difficulty of applying inductive biases, OCL for single-view images remains challenging, resulting in inconsistent learning of object-centric representation. To this end, we introduce a novel OCL framework for single-view images, SLot Attention via SHepherding (SLASH), which consists of two simple-yet-effective modules on top of Slot Attention. The new modules, Attention Refining Kernel (ARK) and Intermediate Point Predictor and Encoder (IPPE), respectively, prevent slots from being distracted by the background noise and indicate locations for slots to focus on to facilitate learning of objectcentric representation. We also propose a weak semisupervision approach for OCL, whilst our proposed framework can be used without any assistant annotation during the inference. Experiments show that our proposed method enables consistent learning of object-centric representation and achieves strong performance across four datasets. Code is available at https://github.com/objectunderstanding/SLASH.

For agents

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

get_harvested_code_for_paper("2303.17842")
get_code_for_paper("2303.17842")
have("2303.17842")

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

curl https://syntology.ai/api/ran/2303.17842.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.17842.svg)](https://syntology.ai/paper/2303.17842)

Connect an agent — have() is free.