The paper investigates how reinforcement learning agents trained in one cyber simulation can be transferred to other simulators or real environments. It introduces a framework that decouples state alignment from action translation, allowing zero‑shot policy transfer without retraining. Experiments across four cyber platforms show that transferred policies can preserve performance in closely aligned settings and achieve substantial win rates in more divergent environments.
arXiv:2603. 19864v2 Announce Type: replace Abstract: Penetration testing, the practice of simulating cyberattacks to identify vulnerabilities, is a complex sequential decision-making task that is inherently partially observable and features large action spaces.
arXiv:2604. 09523v2 Announce Type: replace Abstract: Training reinforcement-learning agents for cyber defense requires an environment that reflects the operational setting: noisy, partial observations, several defenders coordinating across a network, and an adaptive adversary realized through self-play.
arXiv:2610.00590v1 Announce Type: cross Abstract: An autonomous cyber defender trained with reinforcement learning (RL) is typically tied to the network on which it was trained, limiting its ability...
arXiv:2609.16563v1 Announce Type: cross Abstract: We have developed the MAL Simulator, a cyber operation simulator based on the Meta Attack Language (MAL). The MAL Simulator is intended for decision-...
arXiv:2608. 04317v1 Announce Type: cross Abstract: Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied.
Sim2Signal is a benchmark designed to systematically measure the Sim-to-Real gap in traffic signal control by decomposing it into observation, action, transition, and reward gaps. The study evaluates 18 mitigation methods across 33 gap settings and 10 calibrated networks from five real-world locations, finding that direct transfer degrades performance but mitigation effectiveness varies by network and gap type. The most effective approaches tend to estimate the specific changes caused by each gap rather than relying on domain randomization or invariant representations.
AgentServeSim is a simulation framework designed to model the execution of large language model (LLM) agent programs, capturing cross‑turn key‑value (KV) state retention, successor turn release, and scheduling decisions. Unlike existing simulators that operate on request streams, AgentServeSim treats the entire agent program as a single unit of execution, using a Program Control Block, Program Orchestrator, Retention Plane, and Dispatch Plane to emulate realistic serving dynamics. Validation against real vLLM deployments on two GPU platforms shows mean job completion time errors below 5.5%, and the simulator enables automated policy search that improves mean JCT by up to 2.8% over hand‑written policies. whyItMatters":"The simulator provides a realistic, CPU‑based tool for evaluating and optimizing LLM agent serving policies, achieving high fidelity to real deployments and enabling measurable performance gains."
Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost exclusively against static, heuristic red agents, leaving their robustness against adaptive threats critically understudied. Meanwhile, recent advances in Reinforcement Learning with Verifiable Rewards (RLVR) have improved LLM reasoning, but their integration into cybersecurity remains elusive due to the absence of suitable benchmark environments and interaction datasets.
arXiv:2606. 26575v1 Announce Type: cross Abstract: Complex multi-agent control tasks remain challenging for traditional rule-based and model-based approaches, motivating the adoption of learning-based methods.
arXiv:2608. 12253v1 Announce Type: cross Abstract: Multi-agent reinforcement learning for human-AI interaction typically relies on a single large language model to simulate user behavior.
arXiv:2606. 24601v1 Announce Type: new Abstract: Multi-agent reinforcement learning (MARL) addresses the problem of training multiple agents that pursue collaborative, competitive, or mixed objectives.
arXiv:2606. 08168v1 Announce Type: cross Abstract: Leading commercial endpoint detection and response (EDR) products have shifted from operator-configured rule sets to multi-component systems where autonomous AI components operate alongside, and increasingly in place of, operator-deployed policies.