SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.13573 · 2022

Unsupervised Learning of Temporal Abstractions with Slot-based Transformers

arXiv · PDF · Open in the Atlas

Code that ran

7 of 8 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 6 have no contract verdict; 0 failures 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/2203.13573.json · badge: /api/ran/2203.13573.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
agopal42/slottar canonical (classifier label)
classifier basis: plan_row
7 ran · 1 no recorded run
FunctionStatusWhere it lives
build_seq Ran
ran on a synthesized input
agopal42/slottar/slottar_modules.py
code served (permissive licence) · get_code("2fde1372d5034b2e")
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 3, dtype family float, shape (1, 8, 1) 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

def build_seq(seq_length):
    seq_pos = tf.linspace(0.0, 1.0, num=seq_length)
    seq_pos = tf.expand_dims(tf.expand_dims(seq_pos, axis=0), axis=-1)
    seq_pos = tf.cast(seq_pos, dtype=tf.float32)
    return seq_pos

ARGS = {
    "seq_length": 8
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (1, 8, 1) — shape (1, 8, 1) 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
compute_geometric Ran
ran on a synthesized input
agopal42/slottar/utils.py
code served (permissive licence) · get_code("e38b375ee5c1bc1d")
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

# Create a small tensor of lambda values
# lambdas should be shape (batch_size, num_steps) with values in [0, 1]
lambdas_np = np.array([
    [0.1, 0.3, 0.5, 0.7],
    [0.2, 0.4, 0.6, 0.8]
], dtype=np.float32)

lambdas = tf.constant(lambdas_np)

ARGS = {
    "lambdas": lambdas
}
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
create_padding_mask Ran
ran on a synthesized input
agopal42/slottar/slottar_modules.py
code served (permissive licence) · get_code("a40c36a3801e0a48")
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 4, dtype family float, shape (2, 1, 1, 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

# Create test sequences as TensorFlow tensors
# For 'strings' dataset: use 0.0 as padding token
strings_seq = tf.constant([[1.0, 2.0, 3.0, 0.0],
                           [1.0, 0.0, 2.0, 3.0]], dtype=tf.float32)

# For 'craft' dataset: use 6.0 (ompn/compile) or 5.0 (other) as padding token
craft_seq = tf.constant([[1.0, 2.0, 6.0, 3.0],
                         [2.0, 3.0, 1.0, 6.0]], dtype=tf.float32)

# For 'minigrid' dataset: use 7.0 (ompn/compile) or 6.0 (other) as padding token
minigrid_seq = tf.constant([[1.0, 2.0, 7.0, 3.0],
                            [2.0, 3.0, 1.0, 7.0]], dtype=tf.float32)

ARGS = {
    'seq': strings_seq,
    'dataset_id': 'strings',
    'model_type': 'ompn'
}
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 4, dtype family float, shape (2, 1, 1, 4) — shape (2, 1, 1, 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_angles Ran
ran, honoured the contract we drafted
agopal42/slottar/slottar_modules.py
pointer only (licence: MIT); the stored copy get_code serves, from chihchenghsieh/eventlogdice, is pointer only (licence: NONE) · get_code("7816c6d65ab55d8f")
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 chihchenghsieh/eventlogdice, and GitHub detected no licence in that repository when it was checked. 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.

gumbel_sample Ran
ran on a synthesized input
agopal42/slottar/baselines_utils.py
code served (permissive licence) · get_code("98d89b921bd24bf2")
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

def gumbel_sample(shape):
    """Sample Gumbel noise."""
    EPS = 1e-17
    uniform = tf.random.uniform(shape, dtype=tf.dtypes.float32)
    return -tf.math.log(EPS - tf.math.log(uniform + EPS))

ARGS = {
    "shape": (2, 4)
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
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
kl_div Ran
ran on a synthesized input
agopal42/slottar/utils.py
code served (permissive licence) · get_code("58e1bcada748d451")
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 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 tensorflow as tf
import numpy as np

eps = 0.0001

PRIORS = {
    'makeall': np.array([eps, eps, eps, 1.0 - 4 * eps, eps], dtype=np.float32),
    'unlockpickup': np.array([eps, eps, 1.0 - 3 * eps, eps], dtype=np.float32),
    'doorkey': np.array([eps, eps, 1.0 - 3 * eps, eps], dtype=np.float32),
    'blockedunlockpickup': np.array([eps, eps, eps, 0.85, 0.15], dtype=np.float32),
    'keycorridor-s3r3': np.array([eps, eps, eps, 0.3128, 0.2065, 0.3987, 0.0552, 0.0245, 0.002, eps], dtype=np.float32),
    'keycorridor-s4r3': np.array([eps, eps, eps, 0.318, 0.284, 0.267, 0.089, 0.035, 0.008, eps], dtype=np.float32),
    'keycorridor-s6r3': np.array([eps, eps, eps, 0.2414, 0.2841, 0.3012, 0.126, 0.0341, 0.0128, eps], dtype=np.float32)
}

# Create a small valid probability distribution that matches the 'makeall' prior shape
batch_size = 2
num_classes = 5

# Create predictions with proper shape and valid probability values
y_pred_values = np.array([
    [0.0001, 0.0001, 0.0001, 0.9996, 0.0001],
    [0.0001, 0.0001, 0.0001, 0.9996, 0.0001]
], dtype=np.float32)

y_pred = tf.constant(y_pred_values)

dataset_fname = 'makeall'

ARGS = {
    'y_pred': y_pred,
    'dataset_fname': dataset_fname
}
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 0, dtype family float — no output shape is stated: observe_signature.py recorded SymbolicTensor 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
to_one_hot Ran
ran on a synthesized input
agopal42/slottar/baselines_utils.py
code served (permissive licence) · get_code("8d63f384045362b9")
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 (5, 5) 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

# Create tiny test tensors
indices = tf.constant([0, 1, 2, 1, 0], dtype=tf.int32)
max_index = 5

ARGS = {
    "indices": indices,
    "max_index": max_index,
}
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 (5, 5) — shape (5, 5) 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
prepare_dataset Not yet run
unverified (harvested, no recorded run)
agopal42/slottar/preprocess.py
code served (permissive licence) · get_code("ca4ef00b5db7d896")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 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 7 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
get_angles 14 other papers
8 dated earlier, 6 later · 2 only as an assembly brick
1704.04368 (2017-04)
Get To The Point: Summarization with Pointer-Generator Networks
1706.03762 (2017-06)
Attention Is All You Need
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
and 11 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

The discovery of reusable sub-routines simplifies decision-making and planning in complex reinforcement learning problems. Previous approaches propose to learn such temporal abstractions in a purely unsupervised fashion through observing state-action trajectories gathered from executing a policy. However, a current limitation is that they process each trajectory in an entirely sequential manner, which prevents them from revising earlier decisions about sub-routine boundary points in light of new incoming information. In this work we propose SloTTAr, a fully parallel approach that integrates sequence processing Transformers with a Slot Attention module and adaptive computation for learning about the number of such sub-routines in an unsupervised fashion. We demonstrate how SloTTAr is capable of outperforming strong baselines in terms of boundary point discovery, even for sequences containing variable amounts of sub-routines, while being up to 7x faster to train on existing benchmarks.

For agents

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

get_harvested_code_for_paper("2203.13573")
get_code_for_paper("2203.13573")
have("2203.13573")

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

curl https://syntology.ai/api/ran/2203.13573.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.13573.svg)](https://syntology.ai/paper/2203.13573)

Connect an agent — have() is free.