Research map / Adversarial robustness and security / Attacks on text and graphs
TextHoaxer: Budgeted Hard-Label Adversarial Attacks on Text
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Optimize a single adversarial candidate using perturbation-matrix gradients rather than a GA candidate population | Hard-label text attacks · MR, AG, Yahoo, Yelp, IMDB, SNLI, MNLI, mMNLI; budget 1,000 | semantic similarity and perturbation rate: MR: 62.5% -> 67.3% similarity; 14.532% -> 11.905% perturbation rate | Helped |
| Use word-embedding-space optimization with a single candidate | TextHoaxer · IMDB, WordLSTM; varying query budget | semantic similarity and perturbation rate | Helped |
| Use the proposed method instead of baselines | TextHoaxer · Human evaluation of adversarial-example similarity | average rank: 3.39 (SD=1.44) | Mixed |
Closest papers by meaning
- SSPAttack: A Simple and Sweet Paradigm for Black-Box Hard-Label Textual Adversarial Attack (AAAI 2023 · similarity 0.92)
- HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Text (NeurIPS 2023 · similarity 0.91)
- LimeAttack: Local Explainable Method for Textual Hard-Label Adversarial Attack (AAAI 2024 · similarity 0.90)
- Adversarial Paraphrasing: A Universal Attack for Humanizing AI-Generated Text (NeurIPS 2025 · similarity 0.87)
- Revisiting Character-level Adversarial Attacks for Language Models (ICML 2024 · similarity 0.86)
- HQA-VLAttack: Towards High Quality Adversarial Attack on Vision-Language Pre-Trained Models (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.