manas acharya

I am an MSc Artificial Intelligence student with hands-on experience building and evaluating deep learning models for generative modelling of astronomical data, plus project work in model interpretability and fairness-aware learning. I have implemented diffusion and GAN-based architectures, applied explainable AI techniques such as Grad-CAM and LIME, and used feature-space analysis to draw actionable research conclusions.

My applied research includes a Galaxy Classification dissertation that examined class-imbalance mitigation with DDPM, conditional GANs, and hybrid diffusion-GANs, and an ADHD diagnostic support tool leveraging fMRI data with feature selection and demographic-specific optimization to reduce bias. I collaborate closely with supervisors, document experiments rigorously, and iterate on model design and hyperparameters to improve performance and interpretability.