Research Interests
Deep Learning
Reinforcement Learning
Smart Manufacturing
Additive Manufacturing
Research Experience
Boosting Predictive Performance on Tabular Data through Data Augmentation with Latent-Space Flow-Based
Diffusion
Undergraduate Thesis: supervised by Dr. Abdullahil Azeem and Dr. Imtiaz Ahmed
• Investigated the use of flow-based diffusion models for tabular data augmentation in imbalanced datasets.
• Developed ensemble novel data augmentation methods to improve downstream machine learning prediction.
• Implemented and compared various latent-space implementations using autoencoders, PCA and transformers.
• Evaluated the performance of augmented datasets using multiple classifiers and regression models.
• Compared the proposed methods with existing data augmentation techniques.
• Ensured privacy preservation of data in latent space without losing statistical similarity.
• Co-authored a paper on the framework, currently under review by Supervisor.