Research map / Adversarial robustness and security / Attacks on text and graphs
Behavioral Learning in Security Games: Threat of Multi-Step Manipulative Attacks
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 |
|---|---|---|---|
| Manipulate attacks assuming QR when the defender uses QR | Attacker in multi-step security games · 12-target generated security games, T=4 | average attacker utility per step: 0.03 -> 1.03 | Helped |
| Use QR as the assumed defender model when the defender actually uses SUQR | Attacker in multi-step security games · 12-target generated security games, T=4 | average attacker utility per step: 0.03 -> 0.41 | Helped |
| Use QR as the assumed defender model when the defender actually uses SHARP | Attacker in multi-step security games · 12-target generated security games, T=4 | average attacker utility per step: 0.03 -> 0.20 | Helped |
| Use a SHARP surrogate when the defender actually uses QR | Attacker in multi-step security games · 12-target games, T=4 and T=8 | average attacker utility | Hurt |
| Use a mismatched assumed behavior model for attack optimization | Attacker in multi-step security games · Generated security games | average attacker utility | Mixed |
| Extend the horizon from 4 to 8 time steps | Manipulative attacker · Generated security games | average attacker utility per step | Hurt |
| Allow manipulative attacks rather than nonManipulate behavior | Defender facing the attacker · Generated security games | defender utility | Hurt |
Closest papers by meaning
- Manipulating a Learning Defender and Ways to Counteract (NeurIPS 2019 · similarity 0.90)
- Optimal Attack and Defense for Reinforcement Learning (AAAI 2024 · similarity 0.89)
- Rethinking Adversarial Policies: A Generalized Attack Formulation and Provable Defense in RL (ICLR 2024 · similarity 0.85)
- Regret Minimization in Stackelberg Games with Side Information (NeurIPS 2024 · similarity 0.85)
- GUARD: Constructing Realistic Two-Player Matrix and Security Games for Benchmarking Game-Theoretic Algorithms (NeurIPS 2025 · similarity 0.85)
- Feint Behaviors and Strategies: Formalization, Implementation and Evaluation (NeurIPS 2024 · 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.