Research map / Learning theory and algorithms / Submodular and game algorithms
End-to-End Game-Focused Learning of Adversary Behavior in Security Games
Andrew Perrault, Bryan Wilder, Eric Ewing, Aditya Mate, Bistra Dilkina, Milind Tambe
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 |
|---|---|---|---|
| train adversary model end-to-end on defender expected utility instead of two-stage prediction accuracy | GF · Synthetic SSG (8 and 24 targets) | defender expected utility | Improved |
| game-focused vs two-stage on human subject data, 8 targets | GF · Nguyen et al. human data | DEU: 28% more DEU than 2S-GT at 5 attacks | Improved |
| game-focused vs two-stage on human subject data, 24 targets | GF · Nguyen et al. human data | DEU: 66% more DEU than 2S at 1 attack | Improved |
| more training data or fewer features (easier learning) | GF vs 2S-GT · Synthetic SSG | DEU gap | Hurt |
| GF vs 2S-GT on prediction quality | GF · Human subject data | test cross entropy: 2-20% higher cross entropy | Hurt |
| tune two-stage model for small data (2S-GT vs 2S) | two-stage · Synthetic SSG | DEU | Improved |
| GF with less model uncertainty (lower 2S cross entropy) | GF · Synthetic SSG | DEU gap vs 2S | Improved |
Closest papers by meaning
- Manipulating a Learning Defender and Ways to Counteract (NeurIPS 2019 · similarity 0.89)
- Behavioral Learning in Security Games: Threat of Multi-Step Manipulative Attacks (AAAI 2023 · similarity 0.88)
- Learning Optimal Strategies to Commit To (AAAI 2019 · similarity 0.88)
- Regret Minimization in Stackelberg Games with Side Information (NeurIPS 2024 · similarity 0.87)
- Optimal Attack and Defense for Reinforcement Learning (AAAI 2024 · similarity 0.86)
- GUARD: Constructing Realistic Two-Player Matrix and Security Games for Benchmarking Game-Theoretic Algorithms (NeurIPS 2025 · similarity 0.85)
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.
