Regression-aware Continual Learning for Android Malware Detection

2025-07-24

AI Summary

In Android malware detection, continual learning (CL) introduces “security regression”—the re-emergence of previously detected malicious samples as false negatives after model updates—posing severe security risks. Method: This paper formally defines security regression for the first time and proposes a regression-aware continual learning framework. Its core innovation is a model-agnostic Positive Consistency Training (PCT) mechanism, which enforces prediction consistency on historical malicious samples during incremental updates via a regression-aware loss function. Results: Experiments on ELSA, Tesseract, and AZ-Class datasets demonstrate that the method significantly reduces security regression rates while maintaining high detection accuracy and strong evolutionary adaptability, thereby ensuring long-term reliability of malware detection systems.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Multiagent Systems: Multiagent Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Malware evolves rapidly, forcing machine learning (ML)-based detectors to adapt continuously. With antivirus vendors processing hundreds of thousands of new samples daily, datasets can grow to billions of examples, making full retraining impractical. Continual learning (CL) has emerged as a scalable alternative, enabling incremental updates without full data access while mitigating catastrophic forgetting. In this work, we analyze a critical yet overlooked issue in this context: security regression. Unlike forgetting, which manifests as a general performance drop on previously seen data, security regression captures harmful prediction changes at the sample level, such as a malware sample that was once correctly detected but evades detection after a model update. Although often overlooked, regressions pose serious risks in security-critical applications, as the silent reintroduction of previously detected threats in the system may undermine users' trust in the whole updating process. To address this issue, we formalize and quantify security regression in CL-based malware detectors and propose a regression-aware penalty to mitigate it. Specifically, we adapt Positive Congruent Training (PCT) to the CL setting, preserving prior predictive behavior in a model-agnostic manner. Experiments on the ELSA, Tesseract, and AZ-Class datasets show that our method effectively reduces regression across different CL scenarios while maintaining strong detection performance over time.

Problem

• Addressing security regression in continual learning for malware detection
• Mitigating harmful prediction changes in updated malware detectors
• Preserving prior predictive behavior without full retraining

Innovation

• Adapts Positive Congruent Training for CL
• Formalizes and quantifies security regression
• Introduces regression-aware penalty in updates