Papers › Beyond Question-Based Biases: Assessing Multimodal Shortcut Learning in Visual...

Beyond Question-Based Biases: Assessing Multimodal Shortcut Learning in Visual Question Answering

7 Apr 2021ICCV 2021 10arXiv:2104.03149archive 2025-07-28

Corentin Dancette, Remi Cadene, Damien Teney, Matthieu Cord

We introduce an evaluation methodology for visual question answering (VQA) to better diagnose cases of shortcut learning. These cases happen when a model exploits spurious statistical regularities to produce correct answers but does not actually deploy the desired behavior. There is a need to identify possible shortcuts in a dataset and assess their use before deploying a model in the real world. The research community in VQA has focused exclusively on question-based shortcuts, where a model might, for example, answer "What is the color of the sky" with "blue" by relying mostly on the question-conditional training prior and give little weight to visual evidence. We go a step further and consider multimodal shortcuts that involve both questions and images. We first identify potential shortcuts in the popular VQA v2 training set by mining trivial predictive rules such as co-occurrences of words and visual elements. We then introduce VQA-CounterExamples (VQA-CE), an evaluation protocol based on our subset of CounterExamples i.e. image-question-answer triplets where our rules lead to incorrect answers. We use this new evaluation in a large-scale study of existing approaches for VQA. We demonstrate that even state-of-the-art models perform poorly and that existing techniques to reduce biases are largely ineffective in this context. Our findings suggest that past work on question-based biases in VQA has only addressed one facet of a complex issue. The code for our method is available at https://github.com/cdancette/detect-shortcuts.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2104.03149")

Code

Syntology Ran 1 of 2 code samples harvested from 1 repository linked to this paper; 1 has no recorded run. Of those that ran: 1 ran · our draft was wrong.

By repository: official repository: 2 samples from 1 repository, 1 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

cdancette/detect-shortcuts officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

2 samples harvested; 1 ran; 0 honoured the contract we drafted; 1 has no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

1ran · our draft was wrong
1unverified

Licence: 2 of the 2 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from cdancette/detect-shortcuts. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

loadjson cdancette/detect-shortcuts/vqa.py official repository ran · our draft was wrong no licence file found · pointer only · f151d3d0ad479012 · report
run_gminer cdancette/detect-shortcuts/rule_mining.py official repository unverified no licence file found · pointer only · 5d8f3b1c6d1dece9 · report

Tasks

Question AnsweringVisual Question AnsweringVisual Question Answering (VQA)

Datasets

Introduced by this paper, per the archive.

VQA-CE

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Visual Question Answering (VQA) VQA-CE RandImg Accuracy (Counterexamples) 34.41 #1 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE LMH + CSS Accuracy (Counterexamples) 34.36 #2 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE LFF Accuracy (Counterexamples) 34.27 #3 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE LMH Accuracy (Counterexamples) 34.26 #4 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE UpDown Accuracy (Counterexamples) 33.91 #5 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE ESR Accuracy (Counterexamples) 33.26 #6 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE LMH + RMFE Accuracy (Counterexamples) 33.14 #7 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE BLOCK Accuracy (Counterexamples) 32.91 #8 of 9 Archive leaderboard report
Visual Question Answering (VQA) VQA-CE RUBi Accuracy (Counterexamples) 32.25 #9 of 9 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections