Research background / Security & intelligence
Learning defensive strategies against APTs
Reinforcement learning for APT defense
My work includes designing lateral-movement simulation environments and constructing datasets for reinforcement-learning-based APT defense.
This research models attacker decision-making and partial observability, connecting security-domain knowledge with sequential decision-making.
Research approach
- Model lateral movement under partial observability.
- Construct attack simulations and training datasets.
- Investigate proactive defensive strategies.