Alpha claims near-perfect zero-day malware detection by feeding assembly instructions from a dynamic instrumentation tool into a DistilBERT Transformer, but its evaluation fits key classifiers to the test set and relies on training-data overlap.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Zero Day Malware Detection with Alpha: Fast DBI with Transformer Models for Real World Application
Alpha claims near-perfect zero-day malware detection by feeding assembly instructions from a dynamic instrumentation tool into a DistilBERT Transformer, but its evaluation fits key classifiers to the test set and relies on training-data overlap.