Benchmarking Benchmark Leakage in Large Language Models

This repository contains the coding aspect of the project Leakage Detection in Smart Water Distribution Systems, where the collected data is used for training an ANN model to detect leakage.

The AutoMLQuantILDetect package utilizes AutoML approaches to accurately detect and quantify system information leakage. We also provide different approaches to estimate mutual information (MI) within systems that release classification datasets to quantify system information leakage.

Repo contains codes implemented for different side channel attacks and faults for CS579 (physical attacks and countermeasures) for my Master's at Oregon State University

This project was presented at the AWS Impact Plus Hackathon held at Strathmore University. The project was aimed at Water Conservation to counter the effect of Climate change.

Personal Information Tracker (PIT) for Android Applications, Detecting Information Leaks from GUI Components (e.g., EditText, RadioButton, Spinner)

Arduino Home Assistant Water Flowmeter (and Leakage WatchDog)

R package for detecting data leakage in machine learning workflows, ensuring model integrity and reliability.

Capstone A10 | Frontend

Research Artifact of ASIA CCS'25 Paper: *BranchGauge: Modeling and Quantifying Leakage in Randomization-Based Secure Branch Predictors*.

This is me learning how to quickly improve the quality of my models.

This project demonstrates a versatile home automation and safety system using Arduino microcontrollers. It integrates multiple sensors and actuators to provide features like door lock control, temperature monitoring, fire detection, and LPG gas leakage detection.

Detect and prevent data leakage in your datasets using mathematical permutation-based analysis paired with an AI-assisted reporting via an AI agent. Streamlit dashboard for easy inspection, ELI5 explanations, and actionable recommendations.