Research

Defending Stealthy Adversarial Attacks

Multimodal Signal Processing and Synthesis with AI

Human Factor: Cyber-Psychological Behavior Modeling Using Host Data

Insider threat is one of the most damaging threat in enterprise network where a legitimate personnel of an organization uses his/her authorized access in a way that affects the confidentiality, integrity, or availability of information or service. Unintentional or deliberate actions of an enterprise employee can cause insider attack which costs on average $11.45 per threat incident. The frequency of insider threat spiked by 47% in last two years where near about 63% incidents occurs because of employee negligence. According to a report, it takes on average 77 days to detect such threat actors. It is very challenging to detect (un)intentional threat actors analysing the host monitoring data. Therefore, in this research project we are focusing on developing novel Artificial Intelligence (AI) models leveraging deep learning algorithms to push the boundary of insider threat detection strategies. We are incorporating Human Factor (HF) in Cyber behavior analysis to fascilitate extensive analysis of insider threat behavior.

Anomaly-based Intrusion Detection System Design using AI

This project focuses on indentifying novel cyber attacks and root cause analysis of threat actors. We are solving Advanced and Persistent Threat (APT) detection as a graph analysis problem using Graph Neural Network (GNN) because of it’s impressive performance.