Combating Adversarial Misspellings with Robust Word Recognition

2019/05/27 by Danish Pruthi, Bhuwan Dhingra, Pruthi, Danish +3 · 15 citations
Computer Science · #Topic Modeling #Adversarial Robustness in Machine Learning #Hate Speech and Cyberbullying Detection

paper · pdf · doi:10.48550/arxiv.1905.11268

Abstract

To combat adversarial spelling mistakes, we propose placing a word recognition model in front of the downstream classifier. Our word recognition models build upon the RNN semi-character architecture, introducing several new backoff strategies for handling rare and unseen words. Trained to recognize words corrupted by random adds, drops, swaps, and keyboard mistakes, our method achieves 32% relative (and 3.3% absolute) error reduction over the vanilla semi-character model. Notably, our pipeline confers robustness on the downstream classifier, outperforming both adversarial training and off-the-shelf spell checkers. Against a BERT model fine-tuned for sentiment analysis, a single adversarially-chosen character attack lowers accuracy from 90.3% to 45.8%. Our defense restores accuracy to 75%. Surprisingly, better word recognition does not always entail greater robustness. Our analysis reveals that robustness also depends upon a quantity that we denote the sensitivity.

Citations

• ImageNet classification with deep convolutional neural networks

Cited by

• RobustMask: Certified Robustness against Adversarial Neural Ranking Attack via Randomized Masking
• Adversarially Robust Detection of Harmful Online Content: A Computational Design Science Approach
• Governance-Aware Hybrid Fine-Tuning for Multilingual Large Language Models
• Adversarial Attacks and Defense on Texts: A Survey
• BERT-ATTACK: Adversarial Attack Against BERT Using BERT
• Typoglycemia under the Hood: Investigating Language Models' Understanding of Scrambled Words
• Selective Adversarial Attacks on LLM Benchmarks
• Robust ML-based Detection of Conventional, LLM-Generated, and Adversarial Phishing Emails Using Advanced Text Preprocessing
• Evaluating Robustness of Large Language Models Against Multilingual Typographical Errors
• Group-Adaptive Adversarial Learning for Robust Fake News Detection Against Malicious Comments
• SAGE: A Realistic Benchmark for Semantic Understanding
• Human Instruction-Following with Deep Reinforcement Learning via Transfer-Learning from Text
• MALCOM: Generating Malicious Comments to Attack Neural Fake News Detection Models
• Special-Character Adversarial Attacks on Open-Source Language Model
• FireBERT: Hardening BERT-based classifiers against adversarial attack

Related

• BERT-Defense: A Probabilistic Model Based on BERT to Combat Cognitively Inspired Orthographic Adversarial Attacks
• MockingBERT: A Method for Retroactively Adding Resilience to NLP Models
• Robsut Wrod Reocginiton via semi-Character Recurrent Neural Network
• Misspelling Oblivious Word Embeddings
• Misspelling Correction with Pre-trained Contextual Language Model