Questioning the Stability of Visual Question Answering

• URL: http://arxiv.org/abs/2511.11206v1
• Date: Fri, 14 Nov 2025 12:05:05 GMT
• Title: Questioning the Stability of Visual Question Answering
• Authors: Amir Rosenfeld, Neta Glazer, Ethan Fetaya,
• Abstract summary: Visual Language Models (VLMs) have achieved remarkable progress, yet their reliability under small, meaning-preserving input changes remains poorly understood.<n>We present the first large-scale, systematic study of VLM robustness to benign visual and textual perturbations.<n>We show that state-of-the-art systems (e.g., GPT-4o, Gemini 2.0 Flash) frequently fail under shifts as small as a few pixels or harmless rephrasings.
• Score: 11.848401203578456
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Visual Language Models (VLMs) have achieved remarkable progress, yet their reliability under small, meaning-preserving input changes remains poorly understood. We present the first large-scale, systematic study of VLM robustness to benign visual and textual perturbations: pixel-level shifts, light geometric transformations, padded rescaling, paraphrasing, and multilingual rewrites that do not alter the underlying semantics of an image-question pair. Across a broad set of models and datasets, we find that modern VLMs are highly sensitive to such minor perturbations: a substantial fraction of samples change their predicted answer under at least one visual or textual modification. We characterize how this instability varies across perturbation types, question categories, and models, revealing that even state-of-the-art systems (e.g., GPT-4o, Gemini 2.0 Flash) frequently fail under shifts as small as a few pixels or harmless rephrasings. We further show that sample-level stability serves as a strong indicator of correctness: stable samples are consistently far more likely to be answered correctly. Leveraging this, we demonstrate that the stability patterns of small, accessible open-source models can be used to predict the correctness of much larger closed-source models with high precision. Our findings expose a fundamental fragility in current VLMs and highlight the need for robustness evaluations that go beyond adversarial perturbations, focusing instead on invariances that models should reliably uphold.

Related papers

• Same Answer, Different Representations: Hidden instability in VLMs [65.36933543377346]
  We introduce a representation-aware and frequency-aware evaluation framework that measures internal embedding drift, spectral sensitivity, and structural smoothness.<n>We apply this framework to modern Vision Language Models (VLMs) across the SEEDBench, MMMU, and POPE datasets.
  arXiv  Detail & Related papers  (2026-02-06T12:24:26Z)
• Training-Free Generation of Diverse and High-Fidelity Images via Prompt Semantic Space Optimization [50.5332987313297]
  We propose Token-Prompt embedding Space Optimization (TPSO), a training-free and model-agnostic module.<n>TPSO introduces learnable parameters to explore underrepresented regions of the token embedding space, reducing the tendency of the model to repeatedly generate samples from strong modes of the learned distribution.<n>In experiments on MS-COCO and three diffusion backbones, TPSO significantly enhances generative diversity, improving baseline performance from 1.10 to 4.18 points, without sacrificing image quality.
  arXiv  Detail & Related papers  (2025-11-25T00:42:09Z)
• Advancing Reliable Test-Time Adaptation of Vision-Language Models under Visual Variations [67.35596444651037]
  Vision-language models (VLMs) exhibit remarkable zero-shot capabilities but struggle with distribution shifts in downstream tasks when labeled data is unavailable.<n>We propose a Reliable Test-time Adaptation (ReTA) method that enhances reliability from two perspectives.
  arXiv  Detail & Related papers  (2025-07-13T05:37:33Z)
• Seeing is Believing, but How Much? A Comprehensive Analysis of Verbalized Calibration in Vision-Language Models [15.158475816860427]
  Uncertainty is essential for assessing the reliability and trustworthiness of modern AI systems.<n> verbalized uncertainty, where models express their confidence through natural language, has emerged as a lightweight and interpretable solution.<n>However, its effectiveness in vision-language models (VLMs) remains insufficiently studied.
  arXiv  Detail & Related papers  (2025-05-26T17:16:36Z)
• Multimodal LLM-Guided Semantic Correction in Text-to-Image Diffusion [52.315729095824906]
  MLLM Semantic-Corrected Ping-Pong-Ahead Diffusion (PPAD) is a novel framework that introduces a Multimodal Large Language Model (MLLM) as a semantic observer during inference.<n>It performs real-time analysis on intermediate generations, identifies latent semantic inconsistencies, and translates feedback into controllable signals that actively guide the remaining denoising steps.<n>Extensive experiments demonstrate PPAD's significant improvements.
  arXiv  Detail & Related papers  (2025-05-26T14:42:35Z)
• Are vision language models robust to uncertain inputs? [5.249651874118556]
  We show that newer and larger vision language models exhibit improved robustness compared to earlier models, but still suffer from a tendency to strictly follow instructions.<n>For natural images such as ImageNet, this limitation can be overcome without pipeline modifications.<n>We propose a novel mechanism based on caption diversity to reveal a model's internal uncertainty.
  arXiv  Detail & Related papers  (2025-05-17T03:16:49Z)
• FLUKE: A Linguistically-Driven and Task-Agnostic Framework for Robustness Evaluation [24.39952838336609]
  FLUKE is a framework for assessing model robustness through systematic minimal variations of test data.<n>We demonstrate FLUKE's utility by evaluating both fine-tuned models and large language models (LLMs) across six diverse NLP tasks.
  arXiv  Detail & Related papers  (2025-04-24T07:12:37Z)
• Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models? [83.53005932513155]
  Multi-modal large language models (MLLMs) have made significant progress, yet their safety alignment remains limited.<n>We propose finetuning MLLMs on a small set of benign instruct-following data with responses replaced by simple, clear rejection sentences.
  arXiv  Detail & Related papers  (2025-04-14T09:03:51Z)
• Breach in the Shield: Unveiling the Vulnerabilities of Large Language Models [13.216398753024182]
  Large Language Models (LLMs) and Vision-Language Models (VLMs) have achieved impressive performance across a wide range of tasks.<n>In this study, we seek to pinpoint the sources of this fragility by identifying parameters and input dimensions that are susceptible to such perturbations.<n>We propose a stability measure called textbfFI, textbfFirst order local textbfInfluence, which is rooted in information geometry and quantifies the sensitivity of individual parameter and input dimensions.
  arXiv  Detail & Related papers  (2025-03-28T16:23:59Z)
• Rethinking Robustness of Model Attributions [24.317595434521504]
  We show that many attribution methods are fragile and have proposed improvements in either these methods or the model training. We observe two main causes for fragile attributions: first, the existing metrics of robustness over-penalize even reasonable local shifts in attribution. We propose simple ways to strengthen existing metrics and attribution methods that incorporate locality of pixels in robustness metrics and diversity of pixel locations in attributions.
  arXiv  Detail & Related papers  (2023-12-16T20:20:38Z)
• RoAST: Robustifying Language Models via Adversarial Perturbation with Selective Training [105.02614392553198]
  We propose Robustifying LMs via Adversarial perturbation with Selective Training (RoAST) RoAST incorporates two important sources for the model robustness, robustness on the perturbed inputs and generalizable knowledge in pre-trained LMs. We demonstrate the effectiveness of RoAST compared to state-of-the-art fine-tuning methods on six different types of LMs.
  arXiv  Detail & Related papers  (2023-12-07T04:23:36Z)
• Generative Modeling Helps Weak Supervision (and Vice Versa) [87.62271390571837]
  We propose a model fusing weak supervision and generative adversarial networks. It captures discrete variables in the data alongside the weak supervision derived label estimate. It is the first approach to enable data augmentation through weakly supervised synthetic images and pseudolabels.
  arXiv  Detail & Related papers  (2022-03-22T20:24:21Z)
• Understanding the Downstream Instability of Word Embeddings [14.373952177486558]
  Many industrial machine learning (ML) systems require frequent retraining to keep up-to-date with constantly changing data. Small changes in training data can cause significant changes in the model's predictions. We show how a core building block of modern natural language processing pipelines---pre-trained word embeddings---affects the instability of downstream NLP models.
  arXiv  Detail & Related papers  (2020-02-29T00:39:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.