Independent Reading · 2025
Playing Repeated Stochastic Security Games Against Non-Stationary Attackers
Ling Chen and Runfa Zhang · Mathematics
Security GamesNon-StationarityOpponent Modelling
Full text
Read the Paper
Use View PDF or Download PDF to read the full paper on this device.
Paper Information
- Authors
- Ling Chen and Runfa Zhang
- Venue
- Mathematics · 2025
Summary
This paper studies repeated security games where an attacker’s latent behavior type changes stochastically and without notification.
Core Idea
A defender can reuse a library of policies while updating beliefs to detect opponent-type switches and select an appropriate response.
Method
The authors formulate defender strategies for five attacker types, apply Bayesian Policy Reuse, and compare it experimentally with EXP3-S.
Key Takeaways
- Fixed-attacker assumptions can be inadequate for repeated security interactions.
- Belief-based policy reuse can detect behavioral switches and improve defender utility.
- Maintaining a relevant policy and opponent-type library is an important practical dependency.
Questions
- How does the approach respond to attacker behaviors outside the predefined type library?
- Can change detection and policy generation be learned jointly?