# Explainable Autonomous Cyber Defense using Adversarial Multi-Agent Reinforcement Learning

arXiv: https://arxiv.org/abs/2604.04442 · HTML: https://mlsanity.com/paper/2604.04442/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 63 references that are in the map.
- First posted on arXiv 6 Apr 2026.

## Builds on (most cited first, 27 of 27)

- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [Explanation in Artificial Intelligence: Insights from the Social Sciences](https://mlsanity.com/paper/1706.07269/) (2017)
- [Practical Black-Box Attacks against Machine Learning](https://mlsanity.com/paper/1602.02697/) (2016)
- [A Survey and Critique of Multiagent Deep Reinforcement Learning](https://mlsanity.com/paper/1810.05587/) (2018)
- [Counterfactual Explanations and Algorithmic Recourses for Machine Learning: A Review](https://mlsanity.com/paper/2010.10596/) (2020)
- [Deep Reinforcement Learning for Cyber Security](https://mlsanity.com/paper/1906.05799/) (2019)
- [Forewarned is Forearmed: A Survey on Large Language Model-based Agents in Autonomous Cyberattacks](https://mlsanity.com/paper/2505.12786/) (2025)
- [ChatNVD: Advancing Cybersecurity Vulnerability Assessment with Large Language Models](https://mlsanity.com/paper/2412.04756/) (2024)
- [Autonomous Intelligent Cyber-defense Agent (AICA) Reference Architecture. Release 2.0](https://mlsanity.com/paper/1803.10664/) (2018)
- [RegimeFolio: A Regime Aware ML System for Sectoral Portfolio Optimization in Dynamic Markets](https://mlsanity.com/paper/2510.14986/) (2025)
- [A Review on C3I Systems' Security: Vulnerabilities, Attacks, and Countermeasures](https://mlsanity.com/paper/2104.11906/) (2021)
- [Smart HPA: A Resource-Efficient Horizontal Pod Auto-scaler for Microservice Architectures](https://mlsanity.com/paper/2403.07909/) (2024)
- [Towards Deep Learning Enabled Cybersecurity Risk Assessment for Microservice Architectures](https://mlsanity.com/paper/2403.15169/) (2024)
- [Robust Partial Least Squares Using Low Rank and Sparse Decomposition](https://mlsanity.com/paper/2407.06936/) (2024)
- [A Survey on Immersive Cyber Situational Awareness Systems](https://mlsanity.com/paper/2408.07456/) (2024)
- [Machine Learning Driven Smishing Detection Framework for Mobile Security](https://mlsanity.com/paper/2412.09641/) (2024)
- [The Future of AI: Exploring the Potential of Large Concept Models](https://mlsanity.com/paper/2501.05487/) (2025)
- [Co-Evolutionary Defence of Active Directory Attack Graphs via GNN-Approximated Dynamic Programming](https://mlsanity.com/paper/2505.11710/) (2025)
- [Resilient Auto-Scaling of Microservice Architectures with Efficient Resource Management](https://mlsanity.com/paper/2506.05693/) (2025)
- [Hybrid Deep Learning Model using SPCAGAN Augmentation for Insider Threat Analysis](https://mlsanity.com/paper/2203.02855/) (2022)
- [Optimizing Cyber Defense in Dynamic Active Directories through Reinforcement Learning](https://mlsanity.com/paper/2406.19596/) (2024)
- [Microservice Vulnerability Analysis: A Literature Review with Empirical Insights](https://mlsanity.com/paper/2408.03960/) (2024)
- [SCALAR: Self-Calibrating Adaptive Latent Attention Representation Learning](https://mlsanity.com/paper/2510.16474/) (2025)
- [Enhancing Network Resilience through Machine Learning-powered Graph Combinatorial Optimization: Applications in Cyber Defense and Information Diffusion](https://mlsanity.com/paper/2310.10667/) (2023)
- [Unveiling the Black Box: A Multi-Layer Framework for Explaining Reinforcement Learning-Based Cyber Agents](https://mlsanity.com/paper/2505.11708/) (2025)
- [Australian Bushfire Intelligence with AI-Driven Environmental Analytics](https://mlsanity.com/paper/2601.06105/) (2026)

## Built on by (most cited first, 3 of 3)

- [AgenticVM: Agentic AI for Adaptive Software Vulnerability Management](https://mlsanity.com/paper/2605.01739/) (2026)
- [Comparative Analysis of Large Language Models in Healthcare](https://mlsanity.com/paper/2604.10316/) (2026)
- [VACS: Value-Aligned Compositional Shielding for Multi-Agent Reasoning](https://mlsanity.com/paper/2609.26135/) (2026)

## Abstract (arXiv)

Autonomous agents are increasingly deployed in both offensive and defensive cyber operations, creating high-speed, closed-loop interactions in critical infrastructure environments. Advanced Persistent Threat (APT) actors exploit "Living off the Land" techniques and targeted telemetry perturbations to induce ambiguity in monitoring systems, causing automated defenses to overreact or misclassify benign behavior as malicious activity. Existing monolithic and multi-agent defense pipelines largely operate on correlation-based signals, lack structural constraints on response actions, and are vulnerable to reasoning drift under ambiguous or adversarial inputs. We present the Causal Multi-Agent Decision Framework (C-MADF), a structurally constrained architecture for autonomous cyber defense that integrates causal modeling with adversarial dual-policy control. C-MADF first learns a Structural Causal Model (SCM) from historical telemetry and compiles it into an investigation-level Directed Acyclic Graph (DAG) that defines admissible response transitions. This roadmap is formalized as a Markov Decision Process (MDP) whose action space is explicitly restricted to causally consistent transitions. Decision-making within this constrained space is performed by a dual-agent reinforcement learning system in which a threat-optimizing Blue-Team policy is counterbalanced by a conservatively shaped Red-Team policy. Inter-policy disagreement is quantified through a Policy Divergence Score and exposed via a human-in-the-loop interface equipped with an Explainability-Transparency Score that serves as an escalation signal under uncertainty. On the real-world CICIoT2023 dataset, C-MADF reduces the false-positive rate from 11.2%, 9.7%, and 8.4% in three cutting-edge literature baselines to 1.8%, while achieving 0.997 precision, 0.961 recall, and 0.979 F1-score.