Autonomous Multi-agent Cyber Defense: A Novel Approach Using Reinforcement Learning with Hierarchical LLM Critics
Bibliographic record
Abstract
Modern enterprise networks are continuously expanding in both scale and complexity. Alongside this, cyber threats have become more dangerous and dynamic. Consequently, automating cyber tasks by creating AI agents is essential for effectively countering these evolving threats. Reinforcement learning (RL) and Deep Learning (DL) models have shown promise in this area, but suffer from a high false-positive rate, slow convergence, and a lack of context-aware strategies. Incorporation of cyber domain knowledge (threat reports, attack behavior descriptions, LLMs trained on cyber data, etc.) might enable these agents to make informed decisions. In this work, we propose a novel multiagent architecture to enhance RL-based autonomous agents using a hierarchy of large language models (LLMs). Our proposed approach enables real-time adaptation to new attack patterns, potentially without retraining the LLMs. We discuss the use of prompt engineering (to encode organizational policies and shape agent behavior) and retrieval augmented generation to facilitate communication between LLMs and ensure actions are aligned with organizational policies. Our approach aims to bridge semantic understanding with strategic RL-agentic control, offering a scalable and modular solution for autonomous multiagent cyber defense.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".