Applying Machine Learning Methods to Static Analysis of Executables
summary
Describes applying machine learning methods to static analysis of executable files, focusing on using ML techniques to analyze and classify binaries without running them. Covers technologies and approaches from data science and machine learning (e.g., feature extraction from executables, model-based detection or classification) applied to malware or binary analysis tasks.
AI-written summary · about
chapters (21)
- 0:00ML and static analysis
- 2:51Introduction
- 4:37Problem-solving approach
- 6:24Malware file classification
- 8:51Malicious executable features
- 10:39MITRE matrix
- 12:05CAPA static analysis tool
- 14:31Data labeling needs
- 15:56VirusTotal data collection
- 17:41Feature table formation
- 18:23ML model training
- 19:25CAPA limitations
- 22:37Data labeling results
- 24:04Malware class distribution
- 26:55Classification results
- 28:44Imbalanced data impact
- 30:13Data cleaning and accuracy
- 31:58Family classification results
- 33:25Final classifier accuracy
- 34:51Conclusions
- 35:13Q&A
related talks
27mA Machine Learning approach for detecting a Malware
32mAI/Machine Learning: What is Actually Working in Cybersecurity?
30mData Driven Security
53mSecurity of Machine Learning Systems
58mMachine‑assisted AppSec — or how to fire people..
53mMachine Learning to Ultimately Defeat Advanced Ransomware Threats
16mMachine Learning--Cybersecurity Boon or Boondoggle
01h 03mThe Dark Side of Malware Analysis