Research map / Adversarial robustness and security / Black-box and transfer attacks
Optimal Attack against Autoregressive Models by Manipulating the Environment
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 LQR instead of the greedy attack policy | White-box attack on synthetic AR(3) · Synthetic AR(3), 50 trials | actual realized cost: 1492 -> 11.5 | Helped |
| Use LQR instead of no attack | White-box attack on synthetic AR(3) · Synthetic AR(3), 50 trials | actual realized cost: 133.9 -> 11.5 | Helped |
| Use MPC+iLQR instead of no attack | White-box attack on US real GNP · US real GNP, 50 trials | actual realized cost: 6.87×10^{-4} -> 3.03×10^{-4} | Helped |
| Use MPC+iLQR instead of the greedy attack policy | White-box attack on US real GNP · US real GNP, 50 trials | actual realized cost: 3.23×10^{-4} -> 3.03×10^{-4} | Helped |
| Use system identification to estimate models before attacking | Black-box attack on nonlinear dynamics · Black-box nonlinear dynamics, 100 trials | actual realized cost: 1.43 -> 4.20 | Hurt |
Closest papers by meaning
- Data Poisoning Attacks against Autoregressive Models (AAAI 2016 · similarity 0.87)
- Adversarial Attacks on Probabilistic Autoregressive Forecasting Models (ICML 2020 · similarity 0.86)
- BackTime: Backdoor Attacks on Multivariate Time Series Forecasting (NeurIPS 2024 · similarity 0.84)
- Rethinking Adversarial Policies: A Generalized Attack Formulation and Provable Defense in RL (ICLR 2024 · similarity 0.81)
- Maximizing the Value of Predictions in Control: Accuracy Is Not Enough (NeurIPS 2025 · similarity 0.81)
- Adversarial Backdoor Attack by Naturalistic Data Poisoning on Trajectory Prediction in Autonomous Driving (CVPR 2024 · similarity 0.80)
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.