SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.02519 · CVPR · 2023

ANetQA: A Large-scale Benchmark for Fine-grained Compositional Reasoning over Untrimmed Videos

Fei Wu, Zhou Zhao, Jun Yu, Zhou Yu, Kui Ren, Jianping Fan, Lixiang Zheng

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; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 1 failure 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/2305.02519.json · badge: /api/ran/2305.02519.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
MILVLG/anetqa-code canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
FunctionStatusWhere it lives
invert_dict Ran
ran; our contract draft was wrong, not the code
MILVLG/anetqa-code/hcrn/DataLoader.py
code served (permissive licence) · get_code("e6dbe2d3b6f0cc87")
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 dict. 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

torch.manual_seed(0)

# invert_dict takes a dictionary and inverts its key-value pairs
# It simply iterates over d.items() and creates a new dict with swapped k,v
# No special attributes or methods are required - just a standard dict

d = {
    "key1": "value1",
    "key2": "value2",
    "key3": "value3"
}

ARGS = {
    "d": d
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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_vocab Ran
ran on a synthesized input
MILVLG/anetqa-code/hcrn/DataLoader.py
code served (permissive licence) · get_code("4eb168a439e7658b")
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 dict. 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 json
import tempfile
import os

# Create a temporary directory and file for the vocab
temp_dir = tempfile.mkdtemp()
vocab_path = os.path.join(temp_dir, 'vocab.json')

# Create a sample vocab dictionary
vocab_data = {
    'question_token_to_idx': {'what': 0, 'is': 1, 'this': 2},
    'answer_token_to_idx': {'yes': 0, 'no': 1},
    'question_answer_token_to_idx': {'what': 0, 'is': 1, 'yes': 2, 'no': 3}
}

# Write the vocab to a temporary file
with open(vocab_path, 'w') as f:
    json.dump(vocab_data, f)

ARGS = {
    'path': vocab_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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
sample_frames Ran
ran on a synthesized input
MILVLG/anetqa-code/clipbert/sample_imgs_clipbert.py
code served (permissive licence) · get_code("4dadb89db8c00cc2")
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 list, 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.

import random
import numpy as np

# Set seed for reproducibility
random.seed(42)
np.random.seed(42)

# Define ARGS dictionary with all parameters for sample_frames
ARGS = {
    'num_frames': 8,
    'vlen': 100,
    'sample': 'rand',
    'fix_start': None
}
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 list, ndim 1, dtype family int — 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, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 1 later) and 0 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
invert_dict 6 other papers
4 dated earlier, 2 later
1803.05268 (2018-03)
Transparency by Design: Closing the Gap Between Performance and Interpretability in Visual Reasoning
1812.01855 (2018-12)
Explainable and Explicit Visual Reasoning over Scene Graphs
2211.03779 (2022-11)
CRIPP-VQA: Counterfactual Reasoning about Implicit Physical Properties via Video Question Answering
and 3 more not named here
sample_frames 2 other papers
1 dated earlier, 1 the same month
2104.00650 (2021-04)
Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval
2305.03347 (2023-05)
A Large Cross-Modal Video Retrieval Dataset with Reading Comprehension
load_vocab 1 other paper
1 dated earlier
2211.03779 (2022-11)
CRIPP-VQA: Counterfactual Reasoning about Implicit Physical Properties via Video Question Answering

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

Building benchmarks to systemically analyze different capabilities of video question answering (VideoQA) models is challenging yet crucial. Existing benchmarks often use non-compositional simple questions and suffer from language biases, making it difficult to diagnose model weaknesses incisively. A recent benchmark AGQA [8] poses a promising paradigm to generate QA pairs automatically from pre-annotated scene graphs, enabling it to measure diverse reasoning abilities with granular control. However, its questions have limitations in reasoning about the finegrained semantics in videos as such information is absent in its scene graphs. To this end, we present ANetQA, a large-scale benchmark that supports fine-grained compositional reasoning over the challenging untrimmed videos from ActivityNet [4]. Similar to AGQA, the QA pairs in ANetQA are automatically generated from annotated video scene graphs. The fine-grained properties of ANetQA are reflected in the following: (i) untrimmed videos with fine-grained semantics; (ii) spatio-temporal scene graphs with fine-grained taxonomies; and (iii) diverse questions generated from fine-grained templates. ANetQA attains 1.4 billion unbalanced and 13.4 million balanced QA pairs, which is an order of magnitude larger than AGQA with a similar number of videos. Comprehensive experiments are performed for state-of-the-art methods. The best model achieves 44.5% accuracy while human performance tops out at 84.5%, leaving sufficient room for improvement.

For agents

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

get_harvested_code_for_paper("2305.02519")
get_code_for_paper("2305.02519")
have("2305.02519")

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

curl https://syntology.ai/api/ran/2305.02519.json

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

[![Syntology run record](https://syntology.ai/api/ran/2305.02519.svg)](https://syntology.ai/paper/2305.02519)

Connect an agent — have() is free.