Research Projects

Software Vulnerability Detection

This project is part of my Ph.D. research, where I have been working on real-world repository vulnerability dataset curation to characterize patterns of vulnerabilities and its implemented fixes to further train learning engines that can aid the detection of unlabeled vulnerabilities in unseen data, often a very expensive task.

Exploring Single Statement Bugs using Code2Vec

This project works with the ManySSTuBs4J dataset published at MSR 2020, which contains 53K pairs of buggy and fix Java codes for Single Statement Bugs (SStuBs). SStuBs can be fixed in one single line and are labeled into 13 categories by the ManySSTuBs4J dataset. I co-authored with Samuel Flint, from the Program Analysis Lab at UNL, to explore generalizable patterns that can be learned by a machine learning model through these 13 categories, to then predict SSTuBs in unlabeled Java files.

Mapping Physiological Signals to Affective Experiences

This project was an engaging part of my Master's program at Northeastern, where I worked under mentoring of Prof. Ostadabbas to learn how linear mappings can be leveraged to predict affective experiences from physiological signals.