Beyond CAGE: Investigating Generalization of Learned Autonomous Network Defense Policies

Advancements in reinforcement learning (RL) have inspired new directions in intelligent automation of network defense. However, many of these advancements have either outpaced their application to network security or have not considered the challenges associated with implementing them in the real-wo...

Full description

Saved in:

Bibliographic Details
Main Authors:	Wolk, Melody, Applebaum, Andy, Dennler, Camron, Dwyer, Patrick, Moskowitz, Marina, Nguyen, Harold, Nichols, Nicole, Park, Nicole, Rachwalski, Paul, Rau, Frank, Webster, Adrian
Format:	Journal Article
Language:	English
Published:	28-11-2022
Subjects:	Computer Science - Cryptography and Security Computer Science - Learning
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Advancements in reinforcement learning (RL) have inspired new directions in intelligent automation of network defense. However, many of these advancements have either outpaced their application to network security or have not considered the challenges associated with implementing them in the real-world. To understand these problems, this work evaluates several RL approaches implemented in the second edition of the CAGE Challenge, a public competition to build an autonomous network defender agent in a high-fidelity network simulator. Our approaches all build on the Proximal Policy Optimization (PPO) family of algorithms, and include hierarchical RL, action masking, custom training, and ensemble RL. We find that the ensemble RL technique performs strongest, outperforming our other models and taking second place in the competition. To understand applicability to real environments we evaluate each method's ability to generalize to unseen networks and against an unknown attack strategy. In unseen environments, all of our approaches perform worse, with degradation varied based on the type of environmental change. Against an unknown attacker strategy, we found that our models had reduced overall performance even though the new strategy was less efficient than the ones our models trained on. Together, these results highlight promising research directions for autonomous network defense in the real world.
DOI:	10.48550/arxiv.2211.15557