SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2206.01718 · ECCV · 2022

A-OKVQA: A Benchmark for Visual Question Answering using World Knowledge

Roozbeh Mottaghi, Kenneth Marino, Apoorv Khandelwal, Dustin Schwenk, Christopher Clark

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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 0 have no contract verdict; 2 failures of our instrument; 0 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/2206.01718.json · badge: /api/ran/2206.01718.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
allenai/aokvqa canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
eval_aokvqa Ran
ran, honoured the contract we drafted
allenai/aokvqa/evaluation/eval_predictions.py
code served (permissive licence) · get_code("fadea4ef5306c033")
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) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# Construct a minimal dataset in the expected format
# The function expects either a list or dict of items with specific keys
dataset_list = [
    {
        'question_id': 'q1',
        'difficult_direct_answer': False,
        'choices': ['cat', 'dog', 'bird', 'fish'],
        'direct_answers': ['cat', 'cat', 'dog'],
        'correct_choice_idx': 0
    },
    {
        'question_id': 'q2',
        'difficult_direct_answer': False,
        'choices': ['red', 'blue', 'green', 'yellow'],
        'direct_answers': ['blue', 'blue', 'blue'],
        'correct_choice_idx': 1
    },
    {
        'question_id': 'q3',
        'difficult_direct_answer': True,
        'choices': ['yes', 'no', 'maybe', 'unknown'],
        'direct_answers': ['yes', 'no', 'yes'],
        'correct_choice_idx': 0
    }
]

# Construct predictions dict
# Keys are question_ids, values are predicted answers
preds = {
    'q1': 'cat',
    'q2': 'blue',
    'q3': 'yes'
}

ARGS = {
    'dataset': dataset_list,
    'preds': preds,
    'multiple_choice': False,
    'strict': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_coco_path Ran
ran; our contract draft was wrong, not the code
allenai/aokvqa/load_aokvqa.py
code served (permissive licence) · get_code("7546bf7192227c45")
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) and returned str. 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

# Analyze the function:
# - split: string parameter used in f-string (e.g., 'train', 'val')
# - image_id: integer parameter used in f-string with format {:012} (zero-padded 12 digits)
# - coco_dir: string parameter used as base directory path

ARGS = {
    'split': 'train',
    'image_id': 123456789,
    'coco_dir': '/data/coco'
}
Setup block drafted by
a language model, Claude Haiku 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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
load_aokvqa Ran
ran; our contract draft was wrong, not the code
allenai/aokvqa/load_aokvqa.py
code served (permissive licence) · get_code("a5fc4f5a85d234b6")
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) and returned list, 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 os
import json
import tempfile
import shutil

# Create a temporary directory to hold the test data
temp_dir = tempfile.mkdtemp()

# Create a minimal valid aokvqa dataset file
dataset_content = [
    {
        "question_id": 1,
        "image_id": 1,
        "question": "What color is the object?",
        "choices": ["red", "blue", "green", "yellow"],
        "correct_choice_idx": 0,
        "rationales": ["It is clearly red"]
    },
    {
        "question_id": 2,
        "image_id": 2,
        "question": "How many objects are there?",
        "choices": ["one", "two", "three", "four"],
        "correct_choice_idx": 1,
        "rationales": ["There are two objects visible"]
    }
]

# Write the dataset file
dataset_file = os.path.join(temp_dir, 'aokvqa_v1p0_val.json')
with open(dataset_file, 'w') as f:
    json.dump(dataset_content, f)

# Define the arguments
ARGS = {
    'aokvqa_dir': temp_dir,
    'split': 'val',
    'version': 'v1p0'
}
Setup block drafted by
a language model, Claude Haiku 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 list, ndim 1, dtype family float — a list 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 2 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
load_aokvqa 1 other paper
1 later
2025.findings-acl.279 (2025 (year only))
MIRe: Enhancing Multimodal Queries Representation via Fusion-Free Modality Interaction for Multimodal Retrieval

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 Visual Question Answering (VQA) task aspires to provide a meaningful testbed for the development of AI models that can jointly reason over visual and natural language inputs. Despite a proliferation of VQA datasets, this goal is hindered by a set of common limitations. These include a reliance on relatively simplistic questions that are repetitive in both concepts and linguistic structure, little world knowledge needed outside of the paired image, and limited reasoning required to arrive at the correct answer. We introduce A-OKVQA, a crowdsourced dataset composed of a diverse set of about 25K questions requiring a broad base of commonsense and world knowledge to answer. In contrast to existing knowledge-based VQA datasets, the questions generally cannot be answered by simply querying a knowledge base, and instead require some form of commonsense reasoning about the scene depicted in the image. We demonstrate the potential of this new dataset through a detailed analysis of its contents and baseline performance measurements over a variety of state-of-the-art vision-language models. Q: What does the man who sits have trouble doing? MC Answers: (a) Riding (b) Breathing (c) Walking (d) Magic Direct Answers: Walking, Walking, … Rationale: The vehicle being used is for people who cannot use their legs properly and need it for assistance in being mobile. Q: What could block the washer's door? MC Answers: (a) Stool (b) Stove (c) Window (d) Sink Direct Answers: Stove, Oven, … Rationale: The washer door is right in front of the range preventing it from opening. Q: How many people will dine at this table? MC Answers: (a) Two (b) One (c) None (d) Five

For agents

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

get_harvested_code_for_paper("2206.01718")
get_code_for_paper("2206.01718")
have("2206.01718")

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

curl https://syntology.ai/api/ran/2206.01718.json

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

[![Syntology run record](https://syntology.ai/api/ran/2206.01718.svg)](https://syntology.ai/paper/2206.01718)

Connect an agent — have() is free.