Research map / Multi-agent systems and games / Cooperative multi-agent RL
Efficient Adversarial Attacks on Online Multi-agent Reinforcement Learning
Guanlin Liu, Lifeng Lai
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 |
|---|---|---|---|
| Use the mixed attack (action plus reward poisoning) instead of action-only d-portion or reward-only eta-gap attack (ablation) | V-learning agents in a 2-agent, 3-state Markov game, Case 1 (target: both Defect everywhere) · Prisoner-style Markov game, K=10^7 episodes | cumulative attack cost and loss | Improved |
| Compare d-portion, eta-gap and mixed attacks when Conditions 1 and 2 hold (ablation) | V-learning agents in the Markov game, Case 2 (cooperate at s1,s2; defect at s3) · Prisoner-style Markov game, K=10^7 episodes | cumulative attack cost and loss | No clear effect |
| Replace the mixed attack (known target policy) with the approximate mixed attack (black-box, learns target policy in an exploration phase) (ablation) | V-learning agents in the three-robot recycling game · Three recycling robots, m=3, S=8, K=10^7 | cumulative attack cost and loss | No clear effect |
| Use the approximate mixed attack with attacker goal of maximizing robot 1 reward versus minimizing robots 2 and 3 rewards | Approximate mixed attack on V-learning · Three recycling robots | cumulative attack cost and loss | No clear effect |
Closest papers by meaning
- Rethinking Adversarial Policies: A Generalized Attack Formulation and Provable Defense in RL (ICLR 2024 · similarity 0.88)
- Optimal Attack and Defense for Reinforcement Learning (AAAI 2024 · similarity 0.88)
- Adversarial Inception Backdoor Attacks against Reinforcement Learning (ICML 2025 · similarity 0.87)
- Wolfpack Adversarial Attack for Robust Multi-Agent Reinforcement Learning (ICML 2025 · similarity 0.87)
- Exploring the Efficacy of Multi-Agent Reinforcement Learning for Autonomous Cyber Defence: A CAGE Challenge 4 Perspective (AAAI 2025 · similarity 0.87)
- AdvEvo-MARL: Shaping Internalized Safety through Adversarial Co-Evolution in Multi-Agent Reinforcement Learning (ICML 2026 · similarity 0.87)
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.
