Research map / Adversarial robustness and security / Attacks on text and graphs
Data Poisoning Attacks on Multi-Task Relationship 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 |
|---|---|---|---|
| use PATOM-optimized injected data instead of random injected data (direct and indirect) (ablation) | MTRL poisoning attack · Sarcos, Landmine, MNIST | averaged NMSE / AUC of target tasks | Helped |
| direct attack instead of indirect attack at same budget (ablation) | PATOM · Sarcos, Landmine, MNIST | degradation of target tasks: about 50% higher averaged NMSE (Sarcos, 30% budget) | Helped |
| indirect attack exploiting task relatedness on MNIST | PATOM · MNIST | AUC of target tasks | No clear effect |
| choice of (Ttar, Tatt) pair: indirect vs random direct ranking flips | PATOM indirect attack · Sarcos | target loss | Mixed |
| increase budget past 0.1 | PATOM direct attack · MNIST | test AUC | Hurt |
| inject 30% poisoned data into attacking tasks | MTRL learned task correlation · Sarcos T2-T3 correlation | correlation: -0.99 -> 0.99 | Mixed |
Closest papers by meaning
- MTL-UE: Learning to Learn Nothing for Multi-Task Learning (ICML 2025 · similarity 0.87)
- Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples (NeurIPS 2026 · similarity 0.85)
- MP-Nav: Enhancing Data Poisoning Attacks against Multimodal Learning (ICML 2025 · similarity 0.84)
- PoisonBench: Assessing Language Model Vulnerability to Poisoned Preference Data (ICML 2025 · similarity 0.84)
- Phantom Transfer: Data Poisoning can Survive Data-Level Defences (NeurIPS 2026 · similarity 0.83)
- Scaling Trends for Data Poisoning in LLMs (AAAI 2025 · 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.