mlsanityData to 14 Oct 2026

Analysing the Robustness of Vision-Language-Models to Common Corruptions

arXiv 2504.13690 ↗First posted 18 Apr 2025cs.CVcited by 16

On the map: Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

Abstract (arXiv)

Vision-language models (VLMs) have demonstrated impressive capabilities in understanding and reasoning about visual and textual content. However, their robustness to common image corruptions remains under-explored. In this work, we present the first comprehensive analysis of VLM robustness across 19 corruption types from the ImageNet-C benchmark, spanning four categories: noise, blur, weather, and digital distortions. We introduce two new benchmarks, TextVQA-C and GQA-C, to systematically evaluate how corruptions affect scene text understanding and object-based reasoning, respectively. Our analysis reveals that transformer-based VLMs exhibit distinct vulnerability patterns across tasks: text recognition deteriorates most severely under blur and snow corruptions, while object reasoning shows higher sensitivity to corruptions such as frost and impulse noise. We connect these observations to the frequency-domain characteristics of different corruptions, revealing how transformers' inherent bias toward low-frequency processing explains their differential robustness patterns. Our findings provide valuable insights for developing more corruption-robust vision-language models for real-world applications.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 35 of its 39 references that are in the map. First posted on arXiv 18 Apr 2025.

Builds on

35 of 39 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022Flamingo: a Visual Language Model for Few-Shot Learning29 Apr

Other references · 34

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  3. 2020Language Models are Few-Shot Learnerscited by 32,419
  4. 2023GPT-4 Technical Reportcited by 18,396
  5. 2022Training language models to follow instructions with human feedbackcited by 15,971
  6. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  7. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  8. 2023Visual Instruction Tuningcited by 7,869
  9. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  10. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,137
  11. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,326
  12. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,512
22 more
  1. 2022Training Compute-Optimal Large Language Modelscited by 3,019
  2. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,877
  3. 2023MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Modelscited by 2,711
  4. 2022Emergent Abilities of Large Language Modelscited by 2,512
  5. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  6. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,984
  7. 2019LXMERT: Learning Cross-Modality Encoder Representations from Transformerscited by 1,553
  8. 2015DeepFool: a simple and accurate method to fool deep neural networkscited by 1,318
  9. 2021Multimodal Few-Shot Learning with Frozen Language Modelscited by 661
  10. 2021Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modelingcited by 653
  11. 2017One pixel attack for fooling deep neural networkscited by 582
  12. 2021LiT: Zero-Shot Transfer with Locked-image text Tuningcited by 577
  13. 2022Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Modelcited by 559
  14. 2021VisualGPT: Data-efficient Adaptation of Pretrained Language Models for Image Captioningcited by 191
  15. 2021MELM: Data Augmentation with Masked Entity Language Modeling for Low-Resource NERcited by 42
  16. 2022Image Difference Captioning with Pre-training and Contrastive Learningcited by 39
  17. 2023Sibling-Attack: Rethinking Transferable Adversarial Attacks against Face Recognitioncited by 34
  18. 2021UniNet: Unified Architecture Search with Convolution, Transformer, and MLPcited by 23
  19. 2023Defending Against Patch-based Backdoor Attacks on Self-Supervised Learningcited by 20
  20. 2023The Best Defense is a Good Offense: Adversarial Augmentation against Adversarial Attackscited by 15
  21. 2022Encoder-Decoder Architecture for Supervised Dynamic Graph Learning: A Surveycited by 14
  22. 2022Tracing the Origin of Adversarial Attack for Forensic Investigation and Deterrencecited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 13.

16 papers.

Sort:

By area:

  1. 2026VLM-RobustBench: A Comprehensive Benchmark for Robustness of Vision-Language Modelscited by 14
  2. 2025RotBench: Evaluating Multimodal Large Language Models on Identifying Image Rotationcited by 4
  3. 2025Diagnosing Corruption-Induced Reliability Failures in Vision-Language Modelscited by 4
  4. 2026RemoteShield: Enable Robust Multimodal Large Language Models for Earth Observationcited by 2
  5. 2026A Semantic Decoupling-Based Two-Stage Rainy-Day Attack for Revealing Weather Robustness Deficiencies in Vision-Language Modelscited by 1
  6. 2026DUALVISION: RGB-Infrared Multimodal Large Language Models for Robust Visual Reasoningcited by 1
  7. 2026SpaceDG: Benchmarking Spatial Intelligence under Visual Degradationcited by 1
  8. 2025When Harmful Content Gets Camouflaged: Unveiling Perception Failure of LVLMs with CamHarmTIcited by 0
  9. 2026RobustVisRAG: Causality-Aware Vision-Based Retrieval-Augmented Generation under Visual Degradationscited by 0
  10. 2026Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Modelscited by 0
  11. 2026Edge Reliability Gap in Vision-Language Models: Quantifying Failure Modes of Compressed VLMs Under Visual Corruptioncited by 0
  12. 2026When a Zero-Shooter Cheats: Improving Age Estimation via Activation Steeringcited by 0
4 more
  1. 2026Exposing Vulnerabilities in Visible-Infrared VLMs: A Unified Geometric Adversarial Framework with Cross-Task Transferabilitycited by 0
  2. 2026A Mechanistic Analysis of Adversarial Fine-tuning of Vision Transformerscited by 0
  3. 2026Cross-Modal Attention Acts as a Frequency Filter: Why Verbose Prompts Improve Robustness in Vision-Language Modelscited by 0
  4. 2026Later Is Better: Token Reduction for ViTs Under Distribution Shiftcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026BRUCE: Benchmarking Robustness Under Corruption Escalation for Scientific Vision-Language Reasoning
  2. 2026How Robust is OCR-Reasoning? Evaluating OCR-Reasoning Robustness of Vision-Language Models under Visual Perturbations
  3. 2025Evaluating Robustness of Vision-Language Models Under Noisy Conditions
  4. 2024Do Vision-Language Foundational models show Robust Visual Perception?
  5. 2025MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs
  6. 2025On the Domain Robustness of Contrastive Vision-Language Models
  7. 2026Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding?
  8. 2025Unveiling the Lack of LVLM Robustness to Fundamental Visual Variations: Why and Path Forward