# When Vision Fails: Text Attacks Against ViT and OCR

arXiv: https://arxiv.org/abs/2306.07033 · HTML: https://mlsanity.com/paper/2306.07033/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 46 references that are in the map.
- First posted on arXiv 12 Jun 2023.

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [Intriguing properties of neural networks](https://mlsanity.com/paper/1312.6199/) (2013)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [A Call for Clarity in Reporting BLEU Scores](https://mlsanity.com/paper/1804.08771/) (2018)
- [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://mlsanity.com/paper/1904.01038/) (2019)
- [Neural Network Acceptability Judgments](https://mlsanity.com/paper/1805.12471/) (2018)
- [Evasion Attacks against Machine Learning at Test Time](https://mlsanity.com/paper/1708.06131/) (2017)
- [Generating Natural Language Adversarial Examples](https://mlsanity.com/paper/1804.07998/) (2018)
- [TextBugger: Generating Adversarial Text Against Real-world Applications](https://mlsanity.com/paper/1812.05271/) (2018)
- [Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers](https://mlsanity.com/paper/1801.04354/) (2018)
- [Synthetic and Natural Noise Both Break Neural Machine Translation](https://mlsanity.com/paper/1711.02173/) (2017)
- [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://mlsanity.com/paper/2109.10282/) (2021)
- [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://mlsanity.com/paper/2103.06874/) (2021)
- [Scaling Neural Machine Translation](https://mlsanity.com/paper/1806.00187/) (2018)
- [Sponge Examples: Energy-Latency Attacks on Neural Networks](https://mlsanity.com/paper/2006.03463/) (2020)
- [Deceiving Google's Perspective API Built for Detecting Toxic Comments](https://mlsanity.com/paper/1702.08138/) (2017)
- [On the Impact of Various Types of Noise on Neural Machine Translation](https://mlsanity.com/paper/1805.12282/) (2018)
- [Bad Characters: Imperceptible NLP Attacks](https://mlsanity.com/paper/2106.09898/) (2021)
- [Fundamental Tradeoffs between Invariance and Sensitivity to Adversarial Perturbations](https://mlsanity.com/paper/2002.04599/) (2020)
- [Robust Open-Vocabulary Translation from Visual Text Representations](https://mlsanity.com/paper/2104.08211/) (2021)
- [Synthesizing Robust Adversarial Examples](https://mlsanity.com/paper/1707.07397/) (2017)
- [Towards Robust Toxic Content Classification](https://mlsanity.com/paper/1912.06872/) (2019)
- [Fall of Giants: How popular text-based MLaaS fall against a simple evasion attack](https://mlsanity.com/paper/2104.05996/) (2021)
- [FAWA: Fast Adversarial Watermark Attack on Optical Character Recognition (OCR) Systems](https://mlsanity.com/paper/2012.08096/) (2020)

## Built on by (most cited first, 5 of 5)

- [LLM Censorship: A Machine Learning Challenge or a Computer Security Problem?](https://mlsanity.com/paper/2307.10719/) (2023)
- [Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents](https://mlsanity.com/paper/2407.09295/) (2024)
- [Adversarial News and Lost Profits: Manipulating Headlines in LLM-Driven Algorithmic Trading](https://mlsanity.com/paper/2601.13082/) (2026)
- [Vulnerability Analysis of Transformer-based Optical Character Recognition to Adversarial Attacks](https://mlsanity.com/paper/2311.17128/) (2023)
- [IntraGuard: Committee-Side Defenses Against Review Outsourcing to Commercial Chatbots](https://mlsanity.com/paper/2605.05271/) (2026)

## Abstract (arXiv)

Text-based machine learning models are vulnerable to an emerging class of Unicode-based adversarial examples capable of tricking a model into misreading text with potentially disastrous effects. The primary existing defense against these attacks is to preprocess potentially malicious text inputs using optical character recognition (OCR). In theory, OCR models will ignore any malicious Unicode characters and will extract the visually correct input to be fed to the model. In this work, we show that these visual defenses fail to prevent this type of attack. We use a genetic algorithm to generate visual adversarial examples (i.e., OCR outputs) in a black-box setting, demonstrating a highly effective novel attack that substantially reduces the accuracy of OCR and other visual models. Specifically, we use the Unicode functionality of combining characters (e.g., ~n which combines the characters n and ~) to manipulate text inputs so that small visual perturbations appear when the text is displayed. We demonstrate the effectiveness of these attacks in the real world by creating adversarial examples against production models published by Meta, Microsoft, IBM, and Google. We additionally conduct a user study to establish that the model-fooling adversarial examples do not affect human comprehension of the text, showing that language models are uniquely vulnerable to this type of text attack.