Machine Learning Research Assistant - Thiagarajar College of Engineering
Machine Learning Research Assistant
Thiagarajar College of Engineering | April 2022 – May 2025
Research Advisor: Dr. S. Padmavathi (spmcse@tce.edu)
Focus: Explainable AI, Deep Learning, Fairness in ML, UAV/Drone Classification
Published 3 IEEE conference papers as first author on Artificial Intelligence and Machine Learning topics including drone classification, human activity recognition, and solar PV forecasting. View all publications on the Publications page.
- Researched with Defence Institute of Advanced Technology to develop a novel lightweight CNN architecture for UAV/Drone identification and classification. Independently designed Cross Stage Partial Network reducing computational complexity 5×, inference time 20×, and FLOPs 5× while improving accuracy by 5% and reducing model size by 7 times
- Implemented Shapley value-based game theory approach for feature contribution analysis across 500+ sensor attributes in wearable device applications, enhancing model transparency and interpretability for human activity recognition systems
- Conducted research on Explainable AI using SHAP and LIME techniques to interpret deep learning models, ensuring fairness and bias mitigation in machine learning predictions
- Investigated fairness metrics including Statistical Parity Difference (SPD), Equal Opportunity Difference (EOD), and Theil Index (TI) to assess and improve model fairness across demographic groups