We Know Cyber Threat Inside and Out

What is Deep Core Engine?

Based on ALYac, a security program with 16 million users, and more than 1 billion data values, the automated AI platform provides a model for classifying the best new/variable malware through collaboration between malware analysts and AI experts.

In particular, through the latest AIOps rather than a single model, training algorithms suitable for user needs can be applied to satisfy those of various users.

Deep Core - Main Characteristics

  • Smart

    Advanced AI performance through self-training of real-time detection malware

  • Precise

    Minimize the possibility of false detection by double-checking with verification systems based on years of security program operation know-how

  • Agile

    Real-time response through faster analysis compared to static/dynamic analysis.

  • Reliable

    Provides optimal service for the domestic security environment through networking with major domestic/foreign organizations and companies.

Deep Core Engine - Learning Methods

Deep Core looks inside the malware and trains its unique features in various ways.

  • Binary File Embedding Trains and classifies binary codes by picking distinctions from the overall pattern
  • Instruction
    Sequence Embedding
    Identifies the type and location of malicious actions in the pattern of command sequences.
  • Meta Data Embedding Extracts specific metadata from the malware and trains the value through deep learning.
  • Call Graph Embedding Trains the calling pattern of functions that perform malicious actions by understanding the function structure of the entire malware.

Deep Core Platform - Operation Structure

Each element of the platform configures each stage to a plug-in
to train and deploys a model that can classify malware in a customized way.
In this way, we can faithfully meet the needs of our customers even in short project periods.

In addition, as it provides batch management from configuration management to deployment,
it supports customers' environment compatibility as much as possible from on-premises to cloud environments.