AMSI-Based Detection of Malicious PowerShell Code Using Contextual Embeddings

Keyphrases

• Contextual Embeddings 100%
• PowerShell 100%
• Anti-malware 100%
• Unlabeled Data 37%
• Performance Analysis 12%
• True Positive Rate 12%
• False Positive Rate 12%
• Cyber-attacks 12%
• Deep Learning 12%
• Public Repositories 12%
• Model Use 12%
• Attacker 12%
• Token 12%
• Labeled Data 12%
• Malicious Behavior 12%
• Scripting Languages 12%
• Embedding Space 12%
• Large Corpora 12%
• Microsoft 12%
• Contextualized Word Embeddings 12%
• Cybersecurity Domains 12%
• Configuration Management 12%
• Script Engine 12%
• Task Automation 12%
• Windows 10 12%
• Code Detection 12%
• Powershell Scripts 12%
• Automated Management 12%
• Command Line 12%
• Similar Words 12%

Computer Science

• Antimalware 100%
• Unlabeled Data 75%
• Deep Learning Method 25%
• Performance Analysis 25%
• Collected Data 25%
• Embedding Space 25%
• Cyber Attack 25%
• Word Embedding 25%
• Attackers 25%
• Malicious Activity 25%
• False Positive Rate 25%
• Cybersecurity 25%
• Command-Line 25%
• Configuration Management 25%
• True Positive Rate 25%
• Build Interface 25%
• Configuration Task 25%
• Powershell Script 25%
• Scripting Languages 25%
• Public Repository 25%
• Window Machine 25%