A Natural Language Processing Approach for Instruction Set Architecture Identification

no code implementations • 13 Apr 2022 • Dinuka Sahabandu, Sukarno Mertoguno, Radha Poovendran

Empirical evaluations show that using our byte-level features in ML-based ISA identification results in an 8% higher accuracy than the state-of-the-art features based on byte-histograms and byte pattern signatures.

Malware Detection