Contains pipeline tasks for Azure DevOps

The HSA is a statistics-based anomaly detection model that uses graph theory to analyze data relationships across time and scale. It works on diverse data types—such as cybersecurity, materials testing, and GIS—providing unsupervised detection of unusual patterns while reducing false positives with a multi-filter process.

This code processes Zeek logs from network packet captures to identify anomalies in traffic patterns, potentially signaling malicious activity. It uses DBSCAN clustering with Gower distance for feature extraction and a transformer model to assign anomaly scores, with configurable thresholds for detection.

This code detects Living-Off-the-Land attacks by analyzing Zeek logs from network traffic. After preprocessing, it uses K-Means to label devices, applies a Graph Convolutional Network to generate embeddings, and then clusters these embeddings to identify suspicious patterns.

The purpose of this code is to use Machine Learning to spot unusual patterns in network traffic, which could indicate malicious activity. It does this by applying a Graph Neural Network that builds a graph-based view of the network logs and detects anomalies within the local graph context.

COREII Scout helps analysts collect and analyze cyber attack data by automating report writing. Using NLP for entity extraction, ACE for classification, and an LLM for report generation, it streamlines workflows, improves accuracy, and supports efficient threat analysis.

The code includes modules for generating and collecting network traffic data for WIFI networks or a simulated 5G network and tools for training machine learning models to optimize quality of service.

The Tracer Analysis Toolbox (TAT) is a stand-alone executable software package that focuses on reservoir monitoring, using changes in concentration of thermally degrading tracers between and injection well and a production well to reflect the temperature-time histories along flow paths through the reservoir.

This is an application designed to help catalog and make scientific data stored in network storage searchable and actionable by data scientists.

The code fetches an image, places a button behind the image that detects user touch via Unity cues, converts 3D input to pixel space and estimates the corresponding pixel position.