Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2509.10983 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914036246904832 |
|---|---|
| author | Pham, Mai Vaze, Vikrant Chin, Peter |
| author_facet | Pham, Mai Vaze, Vikrant Chin, Peter |
| contents | Cyber defense operations increasingly require long-term strategic planning under uncertainty and resource constraints. We propose a new use of combinatorial auctions for allocating defensive action bundles in a realistic cyber environment, using host-specific valuations derived from reinforcement learning (RL) Q-values. These Q-values encode long-term expected utility, allowing upstream planning. We train CAFormer, a differentiable Transformer-based auction mechanism, to produce allocations that are approximately incentive-compatible under misreporting. Rather than benchmarking against existing agents, we explore the qualitative and strategic properties of the learned mechanisms. Compared to oracle and heuristic allocations, our method achieves competitive revenue while offering robustness to misreporting. In addition, we find that allocation patterns correlate with adversarial and defensive activity, suggesting implicit alignment with operational priorities. Our results demonstrate the viability of auction-based planning in cyber defense and highlight the interpretability benefits of RL-derived value structures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_10983 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Strategic Cyber Defense via Reinforcement Learning-Guided Combinatorial Auctions Pham, Mai Vaze, Vikrant Chin, Peter Computer Science and Game Theory Cyber defense operations increasingly require long-term strategic planning under uncertainty and resource constraints. We propose a new use of combinatorial auctions for allocating defensive action bundles in a realistic cyber environment, using host-specific valuations derived from reinforcement learning (RL) Q-values. These Q-values encode long-term expected utility, allowing upstream planning. We train CAFormer, a differentiable Transformer-based auction mechanism, to produce allocations that are approximately incentive-compatible under misreporting. Rather than benchmarking against existing agents, we explore the qualitative and strategic properties of the learned mechanisms. Compared to oracle and heuristic allocations, our method achieves competitive revenue while offering robustness to misreporting. In addition, we find that allocation patterns correlate with adversarial and defensive activity, suggesting implicit alignment with operational priorities. Our results demonstrate the viability of auction-based planning in cyber defense and highlight the interpretability benefits of RL-derived value structures. |
| title | Strategic Cyber Defense via Reinforcement Learning-Guided Combinatorial Auctions |
| topic | Computer Science and Game Theory |
| url | https://arxiv.org/abs/2509.10983 |