Ayush Meher | Semi-Supervised Variational Autoencoder and Classifier for Synthetic Data Generation

Ayush Meher

Jersey City

Based in United States

About

I hold a Master’s degree in Global Management with a focus on Data Science from Thunderbird School of Global Management, along with a Bachelor's degree in Computer Science. I am currently pursuing full-time opportunities in Data Science, where I can apply my expertise in Machine Learning, Data Visualization, and Advanced Analytics. My passion lies in transforming data into actionable insights that drive business growth and innovation. With a strong interest in Business Consulting, I aim to bridge the gap between technical solutions and strategic decision-making.

Skills

All projects

Semi-Supervised Variational Autoencoder and Classifier for Synthetic Data Generation

Project links

Skills

About this project

This project uses VAEs and CVAEs in PyTorch to generate synthetic, class-conditional data in a semi-supervised setting. It evaluates data quality with Inception Score and FID, and tests generalization with an SVM classifier. PCA and t-SNE visualize class distribution for model insights.