Cognitive Learning Framework for Automated, Robust, and Transparent Security Monitoring & Reasoning

Abstract

Machine learning technologies are stepping in to help improve cyber resilience. Despite the success in machine learning, there are m,ultiple technical challenges towards intelligent security monitoring systems, such as lack of flexibility to adapt to new attacks of, environments, demanding unreasonable resources, and lack of human-like reasoning. Considering the dynamic nature of security proble,ms, we aim to design robust and transparent cognitive learning that provides a high quality of learning and human-level reasoning fo,r each prediction or decision. To achieve such a cognitive system, we exploit HyperDimensional Computing (HDC) as an alternative par,adigm that mimics important brain functionalities towards high-efficiency and noise-tolerant computation. HDC is motivated by the ob,servation that the human brain operates on high-dimensional data representations. In this white paper, we design, HYSecure, a hyperd,imensional cognitive learning framework for robust, efficient, and transparent security monitoring. We first develop flexible operat,ions supporting brain-like information association, memorization, and attention. Our operations enable dynamic, abstract, and human-,interpretable representation of information in high-dimension. We accordingly develop hyperdimensional learning algorithms that are,precisely designed for cybersecurity applications. Our learning models are based on extremely high-dimension and are surprisingly po,werful in detecting noisy, malicious, or anomalous data. We also design automated cognitive computing algorithms that operate over e,ncoded data and provide human-like reasoning on HYSecure decisions. We will evaluate the effectiveness of our framework on multiple,large-scale systems. Our innovations are expected to provide at least two orders of magnitude higher efficiency, advanced learnabili,ty, and human-like reasoning compared to state-of-the-art solutions.The project abstract is approved for public release.

Document Details

Document Type
DoD Grant Award
Publication Date
Dec 10, 2021
Source ID
N000142212067

Entities

People

  • Mohsen Imani

Organizations

  • Office of Naval Research
  • United States Navy
  • University of California, Irvine

Tags

Fields of Study

  • Computer science

Readers

  • Distributed Systems and Data Platform Development
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Cyber