SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2308.14575 · ICCV · 2023

Referring Image Segmentation Using Text Supervision

Ke Xu, Rynson Lau, Fang Liu, Lihe Zhang, Baocai Yin, Yuhao Liu, Yuqiu Kong, Gerhard Hancke

arXiv · PDF · Open in the Atlas

Code that ran

12 of 17 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, 0 violated it and 1 has no contract verdict; 9 failures of our instrument; 5 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/2308.14575.json · badge: /api/ran/2308.14575.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
fawnliu/tris canonical (classifier label)
classifier basis: census
12 ran · 4 no recorded run
fawnliu/TRIS community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
Upsample Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/model/model_stage2.py
code served (permissive licence) · get_code("ff927ca5579e24a3")
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 (2, 4, 6, 10) 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)

# Create a small input tensor for upsampling
# Shape: (batch=2, channels=4, height=3, width=5)
x = torch.randn(2, 4, 3, 5)

# Define the target size for upsampling
# Using distinct dimensions to make shape probing unambiguous
size = (6, 10)

ARGS = {
    'x': x,
    'size': size
}
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 4, dtype family float, shape (2, 4, 6, 10) — shape (2, 4, 6, 10) 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
average_precision Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
fawnliu/tris/IRNet/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from kjmillerCURIS/SPARC, is pointer only (licence: NONE) · get_code("f52b226226096fdd")
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 kjmillerCURIS/SPARC, 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.

compute_mAP Ran
ran, honoured the contract we drafted
fawnliu/tris/IRNet/utils.py
code served (permissive licence) · get_code("4180f96945c4c956")
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 float64, 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 numpy as np

# Set seed for reproducibility
np.random.seed(0)

# Create small sample data for compute_mAP
# Shape: (batch_size, num_classes)
# batch_size=2, num_classes=4
batch_size = 2
num_classes = 4

# targs: binary target labels (0 or 1) for each sample and class
targs = np.array([
    [1, 0, 1, 0],
    [0, 1, 1, 0]
], dtype=np.float32)

# preds: prediction scores (between 0 and 1) for each sample and class
preds = np.array([
    [0.9, 0.2, 0.8, 0.1],
    [0.3, 0.7, 0.6, 0.4]
], dtype=np.float32)

ARGS = {
    "targs": targs,
    "preds": preds
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float 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
create_logger Ran
ran on a synthesized input
fawnliu/tris/logger.py
code served (permissive licence) · get_code("8cc3c8b342e0cdb1")
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 Logger. 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 sys
import logging
import functools
import tempfile
import os

@functools.lru_cache()
def create_logger(output_dir=None, dist_rank=0, name=''):
    logger = logging.getLogger(name)
    logger.setLevel(logging.DEBUG)
    logger.propagate = False
    if dist_rank == 0:
        console_handler = logging.StreamHandler(sys.stdout)
        console_handler.setLevel(logging.DEBUG)
        logger.addHandler(console_handler)
    return logger

# Create a temporary directory for output_dir parameter
temp_dir = tempfile.mkdtemp()

ARGS = {
    'output_dir': temp_dir,
    'dist_rank': 0,
    'name': 'test_logger'
}
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 Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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
criterion Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/model/model_stage2.py
code served (permissive licence) · get_code("bf2dd3002bd9481d")
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 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 torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for the function
# seg_final_out1: logits tensor, shape (batch_size, num_classes) or similar
# target: binary target tensor with same shape as seg_final_out1
batch_size = 2
num_features = 4

seg_final_out1 = torch.randn(batch_size, num_features)
target = torch.randint(0, 2, (batch_size, num_features))

ARGS = {
    'seg_final_out1': seg_final_out1,
    'target': target
}
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 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
de_interleave Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/IRNet/torch_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from kunzhan/InfoMatch, is pointer only (licence: NONE) · get_code("27ae5d495ffa4282")
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 kunzhan/InfoMatch, 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.

dice_loss Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/loss/seg_loss.py
code served (permissive licence) · get_code("e93b4ea79270bbd0")
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 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 torch.nn as nn

torch.manual_seed(0)

# Create inputs: predictions (logits) - shape (batch_size, num_features)
# batch_size=2, num_features=4
inputs = torch.randn(2, 4, dtype=torch.float32)

# Create targets: binary labels - same shape as inputs
targets = torch.randint(0, 2, (2, 4), dtype=torch.float32)

ARGS = {
    "inputs": inputs,
    "targets": targets,
}
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 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
focal_loss Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/model/model_stage1.py
code served (permissive licence) · get_code("cb29cf6c7fa94676")
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

torch.manual_seed(0)

# x should be a tensor that focal_loss will apply operations on
# The function uses: torch.pow(1 - x, p) and torch.log(c + x)
# So x should be a float tensor with values in a reasonable range
# (between 0 and 1 works well for probability-like inputs)
x = torch.tensor([[0.1, 0.5, 0.9, 0.3],
                   [0.2, 0.8, 0.4, 0.6]], dtype=torch.float32)

# p is the power parameter used in torch.pow(1 - x, p)
p = 2

# c is a small constant added in torch.log(c + x) for numerical stability
c = 0.1

ARGS = {
    "x": x,
    "p": p,
    "c": c,
}
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
interleave Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/IRNet/torch_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from kunzhan/InfoMatch, is pointer only (licence: NONE) · get_code("f6043e7798bcabbd")
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 kunzhan/InfoMatch, 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.

pil_loader Ran
ran, honoured the contract we drafted
fawnliu/tris/IRNet/imutils.py
code served (permissive licence) · get_code("4f8ddb114d286a0f")
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 Image. 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 tempfile
import os
from PIL import Image

# Create a temporary directory and a small test image
temp_dir = tempfile.mkdtemp()
test_image_path = os.path.join(temp_dir, "test_image.png")

# Create a small RGB image and save it
test_image = Image.new('RGB', (4, 3), color=(255, 0, 0))
test_image.save(test_image_path)

ARGS = {
    'path': test_image_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
rotation Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/IRNet/torch_utils.py
code served (permissive licence) · get_code("64bd96dab205fe39")
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 (2, 3, 4, 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 torch

torch.manual_seed(0)

# Create a small 4D tensor for rotation
# Shape: (batch=2, height=4, width=3, channels=5)
# Using distinct dimensions so we can verify the operation
x = torch.randn(2, 4, 3, 5)

# k is the number of 90-degree rotations
k = 1

ARGS = {
    "x": x,
    "k": k,
}
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 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 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
sigmoid_focal_loss Ran
ran; our contract draft was wrong, not the code
fawnliu/tris/loss/seg_loss.py
code served (permissive licence) · get_code("9cf331c90b5daee7")
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 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

torch.manual_seed(0)

# Create small tensors for inputs and targets
# inputs: predictions (logits), shape (batch_size, num_classes)
# targets: binary labels, shape (batch_size, num_classes)
batch_size = 2
num_classes = 4

inputs = torch.randn(batch_size, num_classes)
targets = torch.randint(0, 2, (batch_size, num_classes)).float()

ARGS = {
    'inputs': inputs,
    'targets': targets,
    'alpha': 0.25,
    'gamma': 2.0
}
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 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
applyColorMap Not yet run
unverified (harvested, no recorded run)
fawnliu/tris/IRNet/imutils.py
code served (permissive licence) · get_code("2adb54669bb464b5")
How it ran

no invocation recorded.

bilateral_prompt Not yet run
unverified (harvested, no recorded run)
fawnliu/TRIS/model/attn.py
code served (permissive licence) · get_code("bf7bcb3d2c9cf00d")
How it ran

no invocation recorded.

clip_forward Not yet run
unverified (harvested, no recorded run)
fawnliu/tris/loss/clip_loss.py
code served (permissive licence) · get_code("8f142519435534f7")
How it ran

no invocation recorded.

convert_mxnet_to_torch Not yet run
unverified (harvested, no recorded run)
fawnliu/tris/IRNet/utils.py
code served (permissive licence) · get_code("95dfca01d7656b7c")
How it ran

no invocation recorded.

tensor2image Not yet run
unverified (harvested, no recorded run)
fawnliu/tris/IRNet/imutils.py
code served (permissive licence) · get_code("c6139a8991439dca")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 7 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 1 the same month, 5 later) and 10 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
pil_loader 66 other papers
31 dated earlier, 4 the same month, 28 later, 3 same year, order not determinable
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1710.10196 (2017-10)
Progressive Growing of GANs for Improved Quality, Stability, and Variation
1711.02512 (2017-11)
Fine-tuning CNN Image Retrieval with No Human Annotation
and 63 more not named here
average_precision 7 other papers
3 dated earlier, 3 later, 1 same year, order not determinable
2009.14119 (2020-09)
Asymmetric Loss For Multi-Label Classification
2111.12933 (2021-11)
ML-Decoder: Scalable and Versatile Classification Head
2112.07368 (2021-12)
Simple and Robust Loss Design for Multi-Label Learning with Missing Labels
and 4 more not named here
convert_mxnet_to_torch 7 other papers
7 dated earlier
1803.10464 (2018-03)
Learning Pixel-level Semantic Affinity with Image-level Supervision for Weakly Supervised Semantic Segmentation
1903.00179 (2019-03)
Pyramid Feature Attention Network for Saliency detection
1911.01370 (2019-11)
Self-Supervised Difference Detection for Weakly-Supervised Semantic Segmentation
and 4 more not named here
interleave 3 other papers
1 dated earlier, 2 later · 1 only as an assembly brick
2001.07685 (2020-01)
FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
2404.11003 (2024-04)
InfoMatch: Entropy Neural Estimation for Semi-Supervised Image Classification
2405.11756 (2024-05)
Erasing the Bias: Fine-Tuning Foundation Models for Semi-Supervised Learning
de_interleave 2 other papers
1 dated earlier, 1 later · 1 only as an assembly brick
2001.07685 (2020-01)
FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
2404.11003 (2024-04)
InfoMatch: Entropy Neural Estimation for Semi-Supervised Image Classification
Upsample 1 other paper
1 later
2403.02689 (2024-03)
Deep Common Feature Mining for Efficient Video Semantic Segmentation
focal_loss 1 other paper
1 dated earlier
2005.08104 (2020-05)
Single-Stage Semantic Segmentation from Image Labels

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

Existing Referring Image Segmentation (RIS) methods typically require expensive pixel-level or box-level annotations for supervision. In this paper, we observe that the referring texts used in RIS already provide sufficient information to localize the target object. Hence, we propose a novel weakly-supervised RIS framework to formulate the target localization problem as a classification process to differentiate between positive and negative text expressions. While the referring text expressions for an image are used as positive expressions, the referring text expressions from other images can be used as negative expressions for this image. Our framework has three main novelties. First, we propose a bilateral prompt method to facilitate the classification process, by harmonizing the domain discrepancy between visual and linguistic features. Second, we propose a calibration method to reduce noisy background information and improve the correctness of the response maps for target object localization. Third, we propose a positive response map selection strategy to generate highquality pseudo-labels from the enhanced response maps, for training a segmentation network for RIS inference. For evaluation, we propose a new metric to measure localization accuracy. Experiments on four benchmarks show that our framework achieves promising performances to existing fully-supervised RIS methods while outperforming state-ofthe-art weakly-supervised methods adapted from related areas. Code is available at https://github.com/fawnliu/TRIS.

For agents

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

get_harvested_code_for_paper("2308.14575")
get_code_for_paper("2308.14575")
have("2308.14575")

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

curl https://syntology.ai/api/ran/2308.14575.json

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

[![Syntology run record](https://syntology.ai/api/ran/2308.14575.svg)](https://syntology.ai/paper/2308.14575)

Connect an agent — have() is free.