Manati: Machine Learning for Threat Intuitive Analysis
ManaTI Project
Machine Learning for Threat Intuitive Analysis
The ManaTI project consists of a front-end web application and a back-end server infrastructure. The web application centralizes all the operations of the analysts and the back-end infrastructure stores the data and runs the algorithms. The main goal of the web application is to provide threat analysts a fast interface and analysis tools to speed up their research.
The goal of the ManaTI project is to develop machine learning techniques to assist an intuitive threat analyst to speed the discovery of new security problems. The machine learning will contribute to the analysis by finding new relationships and inferences. The project will include the development of a web interface for the analyst to interact with the data and the machine learning output.
This project is partially supported by Cisco Systems.
The main functionality of ManaTI is the weblogs table. This is the structure that holds all the data from the session and where most of the interaction is done. The weblogs table consists of a fast and dynamic JavaScript table of all the weblogs in the session of the analyst. The table is very important because is how ManaTI stores all the weblogs in the memory of the web browser. As soon as a session is created, ManaTI stores the weblogs in the table and it does not send them to the back-end server. This is only done upon request of the analyst by pressing the Save button on top of the page.
