SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.05170 · 2018

xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems

arXiv · PDF · Open in the Atlas

Code that ran

12 of 15 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 9 have no contract verdict; 1 failure 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/1803.05170.json · badge: /api/ran/1803.05170.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
recommenders-team/recommenders community (classifier label)
classifier basis not recorded on the link
7 ran · 1 no recorded run
JianzhouZhan/Awesome-RecSystem-Models community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
wangweitong/DL reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 1 no recorded run
shenweichen/DeepCTR reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 1 no recorded run
FunctionStatusWhere it lives
count_in_filelist_items Ran
ran on a synthesized input
JianzhouZhan/Awesome-RecSystem-Models/util/train_model_util_PyTorch.py
code served (permissive licence) · get_code("93cee4fc981be52d")
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 int, ndim 0, dtype family int, 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 tempfile
import os

# Create temporary files with known content for testing
temp_dir = tempfile.mkdtemp()

# Create test files with different numbers of lines
file1_path = os.path.join(temp_dir, "file1.txt")
with open(file1_path, 'w') as f:
    f.write("line1\nline2\nline3\n")

file2_path = os.path.join(temp_dir, "file2.txt")
with open(file2_path, 'w') as f:
    f.write("lineA\nlineB\n")

file3_path = os.path.join(temp_dir, "file3.txt")
with open(file3_path, 'w') as f:
    f.write("single\n")

# Create filelist with file paths (some with whitespace to test strip())
filelist = [
    f"{file1_path}",
    f"  {file2_path}  ",
    file3_path
]

ARGS = {
    "filelist": filelist
}
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 int, ndim 0, dtype family int, 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
filter_by Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/pandas_df_utils.py
code served (permissive licence) · get_code("c52cd7cbf859f8e0")
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 DataFrame. 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 pandas as pd

# Create sample dataframes for testing the filter_by function
df = pd.DataFrame({
    'id': [1, 2, 3, 4, 5],
    'name': ['Alice', 'Bob', 'Charlie', 'David', 'Eve'],
    'value': [10, 20, 30, 40, 50]
})

filter_by_df = pd.DataFrame({
    'id': [2, 4],
    'name': ['Bob', 'David'],
    'value': [20, 40]
})

filter_by_cols = ['id', 'name']

ARGS = {
    'df': df,
    'filter_by_df': filter_by_df,
    'filter_by_cols': filter_by_cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
find_collection Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/cosmos_cli.py
code served (permissive licence) · get_code("6ec04041da9444a4")
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 bool, 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.

class MockCosmosDBClient:
    """Mock pydocumentdb client for testing."""
    
    def QueryCollections(self, database_link, query_spec):
        """Mock QueryCollections method that returns collection results."""
        # Simulate finding a collection based on the query
        query = query_spec.get('query', '')
        parameters = query_spec.get('parameters', [])
        
        # Extract the id parameter value from the query
        id_value = None
        for param in parameters:
            if param.get('name') == '@id':
                id_value = param.get('value')
                break
        
        # For testing, return a collection if id matches 'test_collection'
        if id_value == 'test_collection':
            return [{'id': 'test_collection', 'resourceId': 'collection_resource_id'}]
        else:
            return []


# Create mock client
mock_client = MockCosmosDBClient()

# Define arguments
ARGS = {
    'client': mock_client,
    'dbid': 'test_db',
    'id': 'test_collection'
}
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 bool, 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
get_idx_value_label Ran
ran; our contract draft was wrong, not the code
JianzhouZhan/Awesome-RecSystem-Models/Model/DeepCrossNetwork_PyTorch.py
pointer only (licence: MIT); the stored copy get_code serves, from wangweitong/DL, is pointer only (licence: NONE) · get_code("a5d2f1c9e119342a")
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 wangweitong/DL, 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.

get_idx_value_label Ran
ran on a synthesized input
JianzhouZhan/Awesome-RecSystem-Models/util/train_model_util_PyTorch.py
code served (permissive licence) · get_code("4ade2988b393a443")
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 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
import tempfile
import os

# Create a temporary file with sample data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_filename = temp_file.name

# Write sample data: label + 13 continuous features + 26 categorical features
# Total: 1 + 13 + 26 = 40 fields
lines = [
    "1\t100\t50\t1000\t5\t100000\t50000\t10000\t1000\t100\t10\t50\t100\t500\tcat1\tcat2\tcat3\tcat4\tcat5\tcat6\tcat7\tcat8\tcat9\tcat10\tcat11\tcat12\tcat13\tcat14\tcat15\tcat16\tcat17\tcat18\tcat19\tcat20\tcat21\tcat22\tcat23\tcat24\tcat25\tcat26",
    "0\t200\t75\t2000\t10\t200000\t75000\t20000\t2000\t200\t20\t100\t200\t1000\tcat1\tcat2\tcat3\tcat4\tcat5\tcat6\tcat7\tcat8\tcat9\tcat10\tcat11\tcat12\tcat13\tcat14\tcat15\tcat16\tcat17\tcat18\tcat19\tcat20\tcat21\tcat22\tcat23\tcat24\tcat25\tcat26",
]

for line in lines:
    temp_file.write(line + '\n')
temp_file.close()

# Build feat_dict_ that maps feature indices and categorical values to IDs
feat_dict_ = {}

# Map continuous feature indices (1-13) to unique IDs
for i in range(1, 14):
    feat_dict_[i] = i

# Map categorical feature values to unique IDs
categorical_values = [f'cat{i}' for i in range(1, 27)]
for idx, cat_val in enumerate(categorical_values, start=100):
    feat_dict_[cat_val] = idx

ARGS = {
    'fname': temp_filename,
    'feat_dict_': feat_dict_,
    'shuffle': True,
}
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 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
get_in_filelist_item_num Ran
ran, honoured the contract we drafted
wangweitong/DL/Model/xDeepFM_PyTorch.py
pointer only (licence: NONE) · get_code("626554fd98c853c6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, wangweitong/DL 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.

is_valid_zip Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/download_utils.py
code served (permissive licence) · get_code("a9c34e87754edb05")
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 bool, 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 tempfile
import zipfile
import os

# Create a temporary valid zip file
temp_dir = tempfile.mkdtemp()
valid_zip_path = os.path.join(temp_dir, "valid.zip")
with zipfile.ZipFile(valid_zip_path, 'w') as z:
    z.writestr("test.txt", "test content")

ARGS = {
    "filepath": valid_zip_path
}
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 bool, 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
process_split_ratio Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/split_utils.py
code served (permissive licence) · get_code("fbf236cde152c63b")
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 tuple, ndim 1, 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 math

ARGS = {
    "ratio": 0.7
}
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 tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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
read_clickhistory Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/mind.py
code served (permissive licence) · get_code("6e86b813e6449904")
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 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 os
import tempfile

# Create a temporary directory and file with sample click history data
temp_dir = tempfile.mkdtemp()
temp_filename = "click_history.txt"
temp_filepath = os.path.join(temp_dir, temp_filename)

# Write sample data in the expected format
# Format: _ \t userid \t imp_time \t click \t imps
# where imps is space-separated "item_id-label" pairs
sample_data = """line1\tuser123\t2023-01-01T10:00:00\titem1 item2 item3\titem1-1 item2-0 item3-1
line2\tuser456\t2023-01-01T10:05:00\titem4 item5\titem4-1 item5-1 item6-0"""

with open(temp_filepath, 'w') as f:
    f.write(sample_data)

ARGS = {
    "path": temp_dir,
    "filename": temp_filename,
}
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 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
remove_duplicates Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/covid_utils.py
code served (permissive licence) · get_code("8b17f7c92b0db3cd")
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 DataFrame. 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 pandas as pd
import numpy as np

# Create a small test dataframe with some duplicate entries
df = pd.DataFrame({
    'col1': [1, 2, 2, 3, 4, 4],
    'col2': ['a', 'b', 'b', 'c', 'd', 'd'],
    'col3': [10, 20, 20, 30, 40, 50]
})

# Define columns to check for duplicates
cols = ['col1', 'col2']

ARGS = {
    'df': df,
    'cols': cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
remove_nan Ran
ran on a synthesized input
recommenders-team/recommenders/recommenders/datasets/covid_utils.py
code served (permissive licence) · get_code("6a132eb3e50b4c7f")
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 DataFrame. 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 pandas as pd

# Create a small test DataFrame with some NaN values
df = pd.DataFrame({
    'col1': [1.0, 2.0, np.nan, 4.0, 5.0],
    'col2': ['a', 'b', '', 'd', 'e'],
    'col3': [10, 20, 30, np.nan, 50]
})

cols = ['col1', 'col2']

ARGS = {
    'df': df,
    'cols': cols
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
version_tuple Ran
ran, honoured the contract we drafted
shenweichen/DeepCTR/examples/run_dien.py
code served (permissive licence) · get_code("b4a378e17a8e73f3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned tuple, ndim 1, dtype family int; 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.

ARGS = {
    'version': '1.2.3'
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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 tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
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_idx_value_label Not yet run
unverified (harvested, no recorded run)
wangweitong/DL/Model/xDeepFM_PyTorch.py
pointer only (licence: NONE) · get_code("13a7876a5bb5d755")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, wangweitong/DL 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. No invocation recorded.

get_xy_fd Not yet run
unverified (harvested, no recorded run)
shenweichen/DeepCTR/examples/run_dien.py
code served (permissive licence) · get_code("dde7d5bc564c9516")
How it ran

no invocation recorded.

load_pandas_df Not yet run
unverified (harvested, no recorded run)
recommenders-team/recommenders/recommenders/datasets/covid_utils.py
code served (permissive licence) · get_code("26d562ddfce33be0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 4 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 11 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_idx_value_label 1 other paper
1 dated earlier
1708.05123 (2017-08)
Deep & Cross Network for Ad Click Predictions
get_in_filelist_item_num 1 other paper
1 dated earlier
1708.05123 (2017-08)
Deep & Cross Network for Ad Click Predictions
get_xy_fd 1 other paper
1 later
1804.04950 (2018-04)
DeepFM: An End-to-End Wide & Deep Learning Framework for CTR Prediction
version_tuple 1 other paper
1 later
1804.04950 (2018-04)
DeepFM: An End-to-End Wide & Deep Learning Framework for CTR Prediction

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

Combinatorial features are essential for the success of many commercial models. Manually crafting these features usually comes with high cost due to the variety, volume and velocity of raw data in web-scale systems. Factorization based models, which measure interactions in terms of vector product, can learn patterns of combinatorial features automatically and generalize to unseen features as well. With the great success of deep neural networks (DNNs) in various fields, recently researchers have proposed several DNN-based factorization model to learn both low- and high-order feature interactions. Despite the powerful ability of learning an arbitrary function from data, plain DNNs generate feature interactions implicitly and at the bit-wise level. In this paper, we propose a novel Compressed Interaction Network (CIN), which aims to generate feature interactions in an explicit fashion and at the vector-wise level. We show that the CIN share some functionalities with convolutional neural networks (CNNs) and recurrent neural networks (RNNs). We further combine a CIN and a classical DNN into one unified model, and named this new model eXtreme Deep Factorization Machine (xDeepFM). On one hand, the xDeepFM is able to learn certain bounded-degree feature interactions explicitly; on the other hand, it can learn arbitrary low- and high-order feature interactions implicitly. We conduct comprehensive experiments on three real-world datasets. Our results demonstrate that xDeepFM outperforms state-of-the-art models. We have released the source code of xDeepFM at \url{https://github.com/Leavingseason/xDeepFM}.

For agents

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

get_harvested_code_for_paper("1803.05170")
get_code_for_paper("1803.05170")
have("1803.05170")

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

curl https://syntology.ai/api/ran/1803.05170.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.05170.svg)](https://syntology.ai/paper/1803.05170)

Connect an agent — have() is free.