A Study on Constructing Knowledge Graph and Graph-based Deep Learning for Prediction and Ranking Problems in Cybersecurity
Abstract
This research aims at studying how to build knowledge graph representation for cybersecurity by performing open information extraction techniques on a large scale of text documents. After that, we will investigate how deep learning can be applied for knowledge graph representation in cybersecurity. We will consider the use of attention mechanism in graph deep learning model and compare various deep learning models which can be used for learning from graph representation data. Besides, we will study semi-supervised learning frameworks for utilizing unlabeled data for improving the performance of prediction and ranking problems. The proposed method will be applied for detecting malware and code flaws in cybersecurity.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 31, 2022
- Accession Number
- AD1166641
Entities
People
- Le-Minh Nguyen
Organizations
- Japan Advanced Institute of Science and Technology