Research map / Adversarial robustness and security / Attacks on text and graphs
Attacks on Online Learners: a Teacher-Student Analysis
Riccardo Giuseppe Margiotta, Sebastian Goldt, Guido Sanguinetti
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| raise attack strength (lower action cost C) past a critical value | Online learner under label attack · Logistic regression, MNIST | student accuracy | Hurt |
| use greedy attack instead of clairvoyant attack | Greedy attacker · Teacher-student perceptron/NN | running cost | Mixed |
| use constant attack instead of greedy attack | Constant attacker · Teacher-student setup | running cost | Hurt |
| decrease batch size for greedy attacks | Greedy attacker · Linear teacher-student | student-target distance | Improved |
| increase batch size P | Greedy attacker · Linear teacher-student | accuracy transition sharpness | Mixed |
| use RL attacker instead of greedy attacker | RL (TD3) attacker · Teacher-student NN | computational cost | Hurt |
| use full training versus transfer learning for the attacked learner | Greedy attack · MNIST and CIFAR10 binary tasks | accuracy vs attack strength | No clear effect |
| add weight decay in transfer learning (VGG11, ResNet18) | Attacked student · CIFAR10 cats vs dogs | student-teacher distance | Mixed |
| swap SGD for Adam in student training | Attacked student · Real-data experiments | learning dynamics | No clear effect |
Closest papers by meaning
- Online Adversarial Attacks (ICLR 2022 · similarity 0.87)
- On the Adversarial Risk of Test Time Adaptation: An Investigation into Realistic Test-Time Data Poisoning (ICLR 2025 · similarity 0.85)
- Practical Adversarial Attacks on Stochastic Bandits via Fake Data Injection (NeurIPS 2026 · similarity 0.85)
- Adversaries With Incentives: A Strategic Alternative to Adversarial Robustness (ICLR 2025 · similarity 0.84)
- Scaling Laws for Adversarial Attacks on Language Model Activations and Tokens (ICLR 2025 · similarity 0.84)
- Adversarial Machine Unlearning (ICLR 2025 · similarity 0.84)
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.
