Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a hybrid deep learning framework for few-shot malware detection, combining an Autoencoder Feature Extractor (AFE) with Model-Agnostic Meta-Learning (MAML). The AFE generates compressed latent representations to mitigate noise and dimensionality, enabling the MAML classifier to rapidly adapt to novel threats with minimal labeled data. Experimental results on the Ransomware Dataset 2024 demonstrate high accuracy, F1 score, and MCC across one to fifty shot settings, highlighting the model's robustness in data-scarce environments.
A novel hybrid deep learning approach achieves robust few-shot malware detection by leveraging autoencoder-derived features for rapid meta-learning adaptation.
Ransomware has emerged as a major cybersecurity threat, with incidents increasing in frequency and impact across critical sectors. These attacks are typically launched through phishing emails, malicious downloads, or exploitation of software vulnerabilities to gain system access. Once inside, the malware encrypts files and demands a ransom, often in cryptocurrency, for the decryption key. Conventional detection methods often struggle with novel or scarce samples, leaving systems vulnerable. To address these challenges, this paper proposes a hybrid deep learning framework that combines an Autoencoder Feature Extractor (AFE) with a Model Agnostic Meta Learning (MAML) classifier for few shot malware detection. The AFE generates compact latent features that reduce noise and dimensionality, while the MAML classifier rapidly adapts to new threats using limited labeled data. Experiments conducted on the Ransomware Dataset 2024 demonstrate the effectiveness of the framework in binary classification tasks. Across one to fifty shot settings, the proposed model consistently achieves high accuracy, F1 score, and Matthews Correlation Coefficient values, maintaining reliable classification even under extreme scarcity. These results highlight the model's robustness and effectiveness in adapting to limited data scenarios, demonstrating the potential of combining feature extraction with meta learning to enhance resilience against malware, particularly in sectors such as healthcare, manufacturing, and public infrastructure, where cyberattacks can cause significant operational and financial disruption.