A Hybrid Approach to Malware Detection: Integrating Few-Shot Model-Agnostic Meta-Learning with Autoencoders
A hybrid autoencoder and MAML model detects ransomware from as few as one labeled sample.
The paper proposes a hybrid detector that pairs an Autoencoder Feature Extractor with Model-Agnostic Meta-Learning for few-shot ransomware classification. The autoencoder builds compact latent features, and MAML adapts to new samples with very few labels. On the Ransomware Dataset 2024, binary classification remains strong from one-shot through fifty-shot settings, with high accuracy, F1, and Matthews correlation.
- Combines an autoencoder feature extractor with a MAML classifier.
- Targets few-shot ransomware detection when labeled samples are scarce.
- Evaluated on the Ransomware Dataset 2024 for binary classification.
- Keeps high accuracy, F1, and MCC from one to fifty shots.
From the original article
1 min read at arxiv.orgRansomware 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.
By Emmanuela Andam · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.