SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.15538 · CVPR · 2021

SUTD-TrafficQA: A Question Answering Benchmark and an Efficient Network for Video Reasoning over Traffic Events

Jun Liu, Li Xu, He Huang

arXiv · PDF · Open in the Atlas

Code that ran

6 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures 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/2103.15538.json · badge: /api/ran/2103.15538.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
MarkHershey/arxiv-dl community (classifier label)
classifier basis not recorded on the link
5 ran · 2 no recorded run
saccharomycetes/text-based-traffic-understanding community
Papers with Code archive
1 ran · 1 no recorded run
FunctionStatusWhere it lives
MyDataset Ran
ran on a synthesized input · constructed an object; no result was computed
saccharomycetes/text-based-traffic-understanding/code/kg_eval.py
pointer only (licence: GPL-3.0) · get_code("f2a7b73a8edfbba6")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, saccharomycetes/text-based-traffic-understanding, is not in the permissive set Syntology serves inline. 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.

bytes_to_mb Ran
ran on a synthesized input
MarkHershey/arxiv-dl/src/arxiv_dl/dl_utils.py
code served (permissive licence) · get_code("c52680018836e274")
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 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

def bytes_to_mb(bytes_int: int) -> float:
    """Convert bytes to megabytes.

    Args:
        bytes_int: Number of bytes

    Returns:
        Size in megabytes as float
    """
    return bytes_int / 1024 / 1024


ARGS = {
    "bytes_int": 5242880,
}
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 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
bytes_to_mb_str Ran
ran on a synthesized input
MarkHershey/arxiv-dl/src/arxiv_dl/dl_utils.py
code served (permissive licence) · get_code("5f5f9fd780fd7de2")
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 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 torch

def bytes_to_mb(bytes_int: int) -> float:
    """Convert bytes to megabytes.

    Args:
        bytes_int: Number of bytes

    Returns:
        Size in megabytes as float
    """
    return bytes_int / 1024 / 1024

def bytes_to_mb_str(bytes_int: int) -> str:
    """Convert bytes to megabytes as formatted string.

    Args:
        bytes_int: Number of bytes

    Returns:
        Size in megabytes as formatted string (e.g., "1.23")
    """
    mb = bytes_to_mb(bytes_int)
    return f'{mb:.2f}'

ARGS = {
    "bytes_int": 5242880
}
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 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
filename_from_url Ran
ran on a synthesized input
MarkHershey/arxiv-dl/src/arxiv_dl/dl_utils.py
code served (permissive licence) · get_code("ed2cd9478b7e5964")
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 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
import urllib.parse as urlparse
from typing import Dict, List, Optional, Union

# Test cases for filename_from_url function
# The function extracts filename from URL using os.path.basename and urlparse.urlparse

ARGS = {
    "url": "https://example.com/path/to/file.txt"
}
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 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
normalize_paper_title Ran
ran on a synthesized input
MarkHershey/arxiv-dl/src/arxiv_dl/helpers.py
code served (permissive licence) · get_code("2068451b9dddb6b3")
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 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 string

def normalize_paper_title(title: str) -> str:
    normalized_title = ''
    for char in title:
        if char in string.printable:
            if char in string.ascii_letters or char in string.digits:
                normalized_title += char
            elif char in string.whitespace:
                normalized_title += '_'
            elif char in '?+-':
                normalized_title += char
            else:
                pass
    return normalized_title

ARGS = {
    'title': 'Machine Learning for NLP: A Survey?'
}
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 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
valid_arxiv_id Ran
ran on a synthesized input
MarkHershey/arxiv-dl/src/arxiv_dl/target_parser.py
code served (permissive licence) · get_code("3bebc05a8d967b01")
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 re

def valid_arxiv_id(paper_id: str) -> bool:
    """
    Validate the arXiv ID according to official arXiv ID format.
    Supports both legacy (pre-2007) and modern (post-2007) arXiv IDs.

    Args: str
        paper_id: arXiv ID to be validated (e.g. '1901.01234', 'math.GT/0309136', 'hep-th/9901001')

    Returns: bool
        True if the arXiv ID is valid. False otherwise.

    Ref: https://arxiv.org/help/arxiv_identifier
    """
    if not isinstance(paper_id, str):
        return False
    modern_pattern = '^(?P<yy>[0-9]{2})(?P<mm>0[1-9]|1[0-2])\\.(?P<num>[0-9]{4,5})(v[0-9]+)?$'
    modern_match = re.fullmatch(modern_pattern, paper_id)
    if modern_match:
        yy = int(modern_match.group('yy'))
        mm = int(modern_match.group('mm'))
        num = modern_match.group('num')
        if yy < 7:
            return False
        if not 1 <= mm <= 12:
            return False
        if yy < 15 or (yy == 14 and mm <= 12):
            if len(num) != 4:
                return False
        elif len(num) != 5:
            return False
        return True
    legacy_pattern = '^(?P<archive>[a-z\\-]+)(\\.[A-Z]{2})?/(?P<ym>[0-9]{4})(?P<seq>[0-9]{3,4})(v[0-9]+)?$'
    legacy_match = re.fullmatch(legacy_pattern, paper_id)
    if legacy_match:
        ym = legacy_match.group('ym')
        seq = legacy_match.group('seq')
        yy = int(ym[:2])
        mm = int(ym[2:])
        if not (91 <= yy <= 99 or 0 <= yy <= 7):
            return False
        if not 1 <= mm <= 12:
            return False
        if len(seq) not in (3, 4):
            return False
        return True
    return False


ARGS = {
    "paper_id": "1901.01234"
}
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
download_paper Not yet run
unverified (harvested, no recorded run)
MarkHershey/arxiv-dl/src/arxiv_dl/__main__.py
code served (permissive licence) · get_code("5409824abd030e72")
How it ran

no invocation recorded.

http_download Not yet run
unverified (harvested, no recorded run)
MarkHershey/arxiv-dl/src/arxiv_dl/helpers.py
code served (permissive licence) · get_code("c42544b64dd4e070")
How it ran

no invocation recorded.

t5_eval Not yet run
unverified (harvested, no recorded run)
saccharomycetes/text-based-traffic-understanding/code/kg_eval.py
pointer only (licence: GPL-3.0) · get_code("b827f2f33d8c9d54")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, saccharomycetes/text-based-traffic-understanding, is not in the permissive set Syntology serves inline. 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.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 9 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.

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

Traffic event cognition and reasoning in videos is an important task that has a wide range of applications in intelligent transportation, assisted driving, and autonomous vehicles. In this paper, we create a novel dataset, SUTD-TrafficQA (Traffic Question Answering), which takes the form of video QA based on the collected 10,080 in-the-wild videos and annotated 62,535 QA pairs, for benchmarking the cognitive capability of causal inference and event understanding models in complex traffic scenarios. Specifically, we propose 6 challenging reasoning tasks corresponding to various traffic scenarios, so as to evaluate the reasoning capability over different kinds of complex yet practical traffic events. Moreover, we propose Eclipse, a novel Efficient glimpse network via dynamic inference, in order to achieve computation-efficient and reliable video reasoning. The experiments show that our method achieves superior performance while reducing the computation cost significantly.

For agents

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

get_harvested_code_for_paper("2103.15538")
get_code_for_paper("2103.15538")
have("2103.15538")

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

curl https://syntology.ai/api/ran/2103.15538.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.15538.svg)](https://syntology.ai/paper/2103.15538)

Connect an agent — have() is free.