SA-DRL: Security-Aware Deep Reinforcement Learning for Ransomware Detection with Asymmetric Reward Design

• URL: http://arxiv.org/abs/2607.06880v1
• Date: Wed, 08 Jul 2026 00:43:20 GMT
• Title: SA-DRL: Security-Aware Deep Reinforcement Learning for Ransomware Detection with Asymmetric Reward Design
• Abstract summary: This study proposes a Security-Aware Deep Reinforcement Learning framework that embeds false-negative and false-positive cost asymmetry into the reinforcement learning reward signal.<n>Four deep reinforcement learning agents were evaluated using a symmetric baseline reward (R1) and a security-aware asymmetric reward (R2)<n>Results show that asymmetric reward shaping improves security-oriented detection performance.
• Score: 1.8024397171920885
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Ransomware detection is a security-critical task in which false negatives and false positives have unequal operational consequences. Conventional machine learning detectors often use symmetric objectives that penalize missed ransomware detections and benign false alarms equally, although a false negative can cause irreversible encryption, operational disruption, and high recovery cost, whereas a false positive is usually reversible. This study proposes a Security-Aware Deep Reinforcement Learning (SA-DRL) framework that embeds false-negative and false-positive cost asymmetry into the reinforcement learning reward signal to prioritize missed-detection reduction. The framework also introduces a Security-Optimal Model Selection (SOMS) criterion and an adaptive episode-level sample-ordering mechanism. Four deep reinforcement learning agents, DQN, DDQN, PPO, and A2C, were evaluated using a symmetric baseline reward (R1) and a security-aware asymmetric reward (R2). Experiments used four discount factors, five-fold cross-validation, and three random seeds, resulting in 480 training runs on a balanced ransomware detection dataset. The SOMS criterion selects models by prioritizing false-negative rate, followed by F1-score and training time. Results show that asymmetric reward shaping improves security-oriented detection performance. The SOMS-selected configuration, DDQN with R2 and gamma = 0.1, achieved a false-negative rate of 0.0080, an F1-score of 0.9915, and an AUC of 0.998, reducing missed detections by 67.6% compared with the best supervised baseline. Across all configurations, R2 reduced the mean false-negative rate by 43% relative to R1. These findings show that reward-function design is important for security-sensitive ransomware detection.

Related papers

• Auditable Machine Unlearning for Privacy-Compliant Ransomware Detection Using Multi-Shard SISA and Deep Reinforcement Learning [1.8024397171920885]
  We propose an auditable ransomware detection and unlearning framework that integrates deep reinforcement learning with multi-shard SISA retraining.<n>In the proposed system, a Double Deep Q-Network (DDQN) learns a reward-guided detection policy from behavioral features under asymmetric security costs.<n>We show that single-shard unlearning maintains minimal utility degradation and low oracle disagreement, whereas moderate shard counts provide the best efficiency-performance trade-off.
  arXiv  Detail & Related papers  (2026-07-07T23:23:52Z)
• Privacy-Aware Machine Unlearning with SISA for Reinforcement Learning-Based Ransomware Detection [1.8024397171920885]
  We present a privacy-aware machine unlearning evaluation framework for ransomware detection built on Sharded, Isolated, Sliced, and Aggregated training.<n>The framework enables efficient data deletion by retraining only the affected model shards rather than the entire detector.
  arXiv  Detail & Related papers  (2026-04-18T00:19:20Z)
• Measuring and Exploiting Confirmation Bias in LLM-Assisted Security Code Review [6.417595678110472]
  We study whether confirmation bias affects LLM-based vulnerability detection, and whether this failure mode can be exploited in software supply-chain attacks.<n>Study 1 quantifies confirmation bias through controlled experiments on 250 CVE vulnerability/patch pairs evaluated across four state-of-the-art models under five framing conditions for the review prompt.<n>Study 2 evaluates exploitability in practice mimicking adversarial pull requests that reintroduce known vulnerabilities while framed as security improvements or urgent functionality fixes via their pull request metadata.
  arXiv  Detail & Related papers  (2026-03-19T10:40:27Z)
• Reward Under Attack: Analyzing the Robustness and Hackability of Process Reward Models [68.45272703833209]
  We show that state-of-the-art PRMs are systematically exploitable under adversarial optimization pressure.<n>We introduce a three-tiered diagnostic framework that applies increasing adversarial pressure to quantify these vulnerabilities.<n>We release PRM-BiasBench and a diagnostic toolkit to enable robustness evaluation before deployment.
  arXiv  Detail & Related papers  (2026-02-20T23:38:03Z)
• StealthRL: Reinforcement Learning Paraphrase Attacks for Multi-Detector Evasion of AI-Text Detectors [0.0]
  StealthRL is a reinforcement learning framework for AI-text detectors.<n>We train a paraphrase policy against a multi-detector ensemble under realistic adversarial conditions.<n>StealthRL achieves near-zero detection, reduces mean AUROC from 0.74 to 0.27, and attains a 99.9% attack success rate.
  arXiv  Detail & Related papers  (2026-02-09T17:33:46Z)
• CVeDRL: An Efficient Code Verifier via Difficulty-aware Reinforcement Learning [57.24524263804788]
  Code verifiers play a critical role in post-verification for LLM-based code generation.<n>Existing supervised fine-tuning methods suffer from data scarcity, high failure rates, and poor inference efficiency.<n>We show that naive RL with only functionality rewards fails to generate effective unit tests for difficult branches and samples.
  arXiv  Detail & Related papers  (2026-01-30T10:33:29Z)
• CARE What Fails: Contrastive Anchored-REflection for Verifiable Multimodal [84.71254539482369]
  Group-relative reinforcement learning with verifiable rewards (RLVR) often wastes the most informative data it already has the failures.<n>We present CARE, a failure-centric post-training framework for multimodal reasoning that turns errors into supervision.<n> CARE improves accuracy and training smoothness while explicitly increasing the share of learning signal that comes from failures.
  arXiv  Detail & Related papers  (2025-12-22T16:34:21Z)
• Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers [90.50039419576807]
  Reinforcement Learning with Verifiable Rewards (RLVR) trains policies against automated verifiers to avoid costly human labeling.<n>To reduce vulnerability to verifier hacking, many RLVR systems collapse rewards to binary $0,1$ during training.<n>This choice carries a cost: it introduces textitfalse negatives (rejecting correct answers, FNs) and textitfalse positives (accepting incorrect ones, FPs)
  arXiv  Detail & Related papers  (2025-10-01T13:56:44Z)
• Lie Detector: Unified Backdoor Detection via Cross-Examination Framework [68.45399098884364]
  We propose a unified backdoor detection framework in the semi-honest setting.<n>Our method achieves superior detection performance, improving accuracy by 5.4%, 1.6%, and 11.9% over SoTA baselines.<n> Notably, it is the first to effectively detect backdoors in multimodal large language models.
  arXiv  Detail & Related papers  (2025-03-21T06:12:06Z)
• Leveraging Uncertainty for Improved Static Malware Detection Under Extreme False Positive Constraints [21.241478970181912]
  We show how ensembling and Bayesian treatments of machine learning methods for static malware detection allow for improved identification of model errors. In particular, we improve the true positive rate (TPR) at an actual realized FPR of 1e-5 from an expected 0.69 for previous methods to 0.80 on the best performing model class on the Sophos industry scale dataset.
  arXiv  Detail & Related papers  (2021-08-09T14:30:23Z)
• Transferable, Controllable, and Inconspicuous Adversarial Attacks on Person Re-identification With Deep Mis-Ranking [83.48804199140758]
  We propose a learning-to-mis-rank formulation to perturb the ranking of the system output. We also perform a back-box attack by developing a novel multi-stage network architecture. Our method can control the number of malicious pixels by using differentiable multi-shot sampling.
  arXiv  Detail & Related papers  (2020-04-08T18:48:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.