Research map / Adversarial robustness and security / Attacks on text and graphs
Contamination Attacks and Mitigation in Multi-Party Machine Learning
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 |
|---|---|---|---|
| Add 5% contaminated records linking Black to low education level | Multi-party model · Adult | contamination accuracy: 0% to 22% | Helped |
| Add 5% contaminated records linking single men to default | Multi-party model · Credit Card | contamination accuracy: 11% to 23% | Helped |
| Add 5% contaminated records linking Computer to Baseball | Multi-party model · News20 | contamination accuracy: 0% to 37% | Helped |
| Add 5% contaminated records linking BMW to Baseball | Multi-party model · News20 | contamination accuracy: 0% to 38% | Helped |
| Adversarially train predictions to obscure party identity | Multi-party model · Adult, Credit Card, and News20 | contamination accuracy | Hurt |
| Adversarially train predictions to obscure party identity | Multi-party model · Adult, Credit Card, and News20 | validation accuracy | Helped |
| Adversarially train predictions to obscure party identity | Multi-party model · Adult with parties partitioned by occupation | validation accuracy: decreases ... by only 0.6% | Hurt |
| Adversarially train predictions to obscure party identity | Multi-party model · Adult, different-distribution parties | party identifier accuracy: 0.642 -> 0.193 | Hurt |
| Adversarially train predictions to obscure party identity | Multi-party model · Adult, different-distribution parties | validation accuracy: 0.715 -> 0.709 | Hurt |
| Use party-identity adversarial training when an attacker controls multiple parties | Multi-party model · Adult, attacker controls up to seven of ten parties | contamination accuracy | Hurt |
| Use adversarial training rather than no adversarial training | Adult multi-party model · Adult, varying contamination fraction | contamination accuracy | Hurt |
Closest papers by meaning
- Adversarial Machine Unlearning (ICLR 2025 · similarity 0.86)
- Reconstruction Attacks on Machine Unlearning: Simple Models are Vulnerable (NeurIPS 2024 · similarity 0.85)
- Can Private Machine Learning Be Fair? (AAAI 2025 · similarity 0.84)
- Privacy Backdoors: Stealing Data with Corrupted Pretrained Models (ICML 2024 · similarity 0.84)
- Rethinking Adversarial Robustness in the Context of the Right to be Forgotten (ICML 2024 · similarity 0.84)
- Privacy Backdoors: Enhancing Membership Inference through Poisoning Pre-trained Models (NeurIPS 2024 · similarity 0.83)
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.